Evaluation of nonpoint-source contamination, Wisconsin; selected streamwater-quality data, land-use and best-management practices inventory, and quality assurance and quality control, water year 1993

Corsi, Steven R. srcorsi@usgs.gov, Walker, John F. jfwalker@usgs.gov, Graczyk, D.J., Greb, S.R. · 1995 · Report

The objective of the watershed-management evaluation monitoring program in Wisconsin is to evaluate the effectiveness of the best-management practices (BMPs)

This 1995 document, Evaluation of nonpoint-source contamination, Wisconsin; selected streamwater-quality data, land-use and best-management practices inventory, and quality assurance and quality control, water year 1993, is preserved in the Mountain Man Mining Library for research and reference. Original source: pubs.usgs.gov.

EVALUATION OF NONPOINT-SOURCE CONTAMINATION, WISCONSIN: SELECTED STREAMWATER-QUALITY DATA, LAND-USE AND BEST-MANAGEMENT PRAC- TICES INVENTORY, AND QUALITY ASSURANCE AND QUALITY CONTROL, WATER YEAR 1993 By S.R. Corsi, J.F. Walker, D.J. Graczyk, S.R. Greb, D.W. Owens, and K.F. Rappc'd U.S. GEOLOGICAL SURVEY Open-File Report 94-707 Prepared in cooperation with the WISCONSIN DEPARTMENT OF NATURAL RESOURCES Madison, Wisconsin

CONTENTS Page Abstract 1 Introduction 1 Selected streamwater-quality data 2 Constituent loads in storm runoff 3 Summary of data 3 Summary of storm-sampling efficiency 4 Monitoring activities in water year 1994 4 Dissolved oxygen 4 Summary of data 9 Monitoring activities in water year 1994 9 Total-recoverable and dissolved hardness 9 Pesticides 11 Summary of data 11 Rural sites 11 Urban sites 17 Monitoring activities in water year 1994 17 Land-use and best-management-practices inventory 20 Data collection 20 Agricultural land use 20 Development of geographic-information-system 21 Activities in water year 1994 21 Quality assurance and quality control 22 Inorganic and organic constituents at urban sites 22 Inorganic constituents 22 Sample bottles 22 Filter-blank samples 22 Splitter-blank samples 24 Pump-tube-blank samples 24 Sampler-blank samples 24 Organic constituents 24 Significance of sample contamination 24 Fecal coliform bacteria 25 Data analysis 26 Application oft-test 26 Variance-component analysis 30 Correlation of fecal-coliform concentrations with other water-quality variables 30 Possible methods to improve analytical results 30 Summary 31 References cited 32 Appendixes: 1. Storm-load data for rural watershed-management evaluation monitoring sites, Wisconsin, water years 1985-93 35 2. Storm-load data for urban watershed-management evaluation monitoring sites, Wisconsin, water years 1991-93 45 3. Quality-assurance/quality-control plan for urban watershed-management evaluation monitoring program, Wisconsin 54 4. Summary of constituent concentrations indicative of contamination in sample blacks for urban watershed-management evaluation monitoring sites, Wisconsin 57

ILLUSTRATIONS Page Figure 1. Map showing location of rural and urban sites in the Wisconsin watersheimanagement evaluation monitoring program, water year 1993 2 2-9. Graphs showing total rainfall and runoff for selected hydrograph-rise periods at: 2. Brewery Creek monitoring site, Dane County, Wis., water year 1993 5 3. Garfoot Creek monitoring site, Dane County, Wis., water year 1993 5 4. Eagle Creek monitoring site, Buffalo County, Wis., water year 1993 6 5. Joos Valley Creek monitoring site, Buffalo County, Wis., water year 1993 6 6. Bower Creek monitoring site, Brown County, Wis., water year 1993 7 7. Otter Creek monitoring site, Sheboygan County, Wis., water year 1993 7 8. Kuenster Creek monitoring site, Grant County, Wis., water year 1993 8 9. Rattlesnake Creek monitoring site, Grant County, Wis., water year 1993 8 10-18. Graphs showing: 10. Return period in days that the dissolved-oxygen concentration was less thrn a given concentration for one continuous hour, summer 1993, at selected coldwater stream sites in Wisconsin 12 11. Return period in days that the dissolved-oxygen concentration was less thrn a given concentration for one continuous hour, summer 1993, at selected warmwater stream sites in Wisconsin 12 12. Relation between total-recoverable hardness and dissolved hardness of stormflow and fixed-interval water samples from the urban watersheimanagement evaluation monitoring sites, Wisconsin, water years 1992 and 1993 13 13. Concentrations of atrazine, alachlor, metolachlor, and cyanazine in water samples at the six rural watershed-management evaluation monitoring sites, Wisconsin, water year 1993 16 14. Concentrations of alachlor, atrazine, and 2,4-D in water samples at the four urban watershed-management evaluation monitoring sites, Wisconsin, water year 1993 19 15. Filter-blank concentrations for selected dissolved constituents for the urban watershed-management evaluation monitoring program, Wisconsin, water year 1993 23 16. Mean fecal coliform colony counts for different holding times of replicate water samples from Brewery Creek, Wis., for two storms in 1993 27 17. Mean fecal coliform colony counts for different holding times of replica4 water samples from Garfoot Creek, Wis., for three storms in 1993 28 18. Examples of the effect of holding time on maximum, mean, and minimum fecal coliform colony counts in water samples from Brewery and Garfcnt Creeks, Wis 29

TABLES Page Table 1. Location, site type, and principal function of sites in the watershed-management evaluation monitoring program, Wisconsin, water year 1993 3 2. Summary of surface water dissolved-oxygen concentration data collected at Wisconsin watershed-management evaluation monitoring sites, water years 1990-93 10 3. Number of days that surface water dissolved-oxygen concentration was less than the State of Wisconsin standard at selected stream sites during water years 1991-93 11 4. Summary of pesticide data in nonfiltered surface water samples from the six rural watershed-management evaluation monitoring sites in Wisconsin, water year 1993 14 5. Summary of pesticide data in nonfiltered surface water samples from the four urban watershed-management evaluation monitoring sites in Wisconsin, water year 1993 18 6. Agricultural land use in selected rural watersheds in Wisconsin 21 7. Fecal coliform summary statistics, Brewery and Garfoot Creeks, Wis., water years 1993-94 26 8. Fecal coliform counts in, and dilutions used for, the water sample collected from Brewery Creek, Wis., on September 13, 1994 30 9. Values of the three major components of variance in fecal coliform concentrations in water samples from Brewery and Garfoot Creeks, Wis., water year 1993 31 10. Pearson correlation coefficients between fecal coliform concentrations in, and selected characteristics of, water samples from Brewery and Garfoot Creeks, Wis., water year 1993 31

CONVERSION FACTORS AND VERTICAL DATUM Multiply By To obtain inch (in.) millimeter foot (ft) meter mile (mi) kilometer acre hectare square mile (mi2) square kilometer million cubic feet (Mft3) million cubic meter? gallon (gal) liter cubic foot per second (ft3/s) cubic meter per second pound (Ib) gram Abbreviated water-quality units used in this report: Chemical concentrations and water temperature are given in metric units. Chemical concentration is given in milligrams per liter (mg/L) or micrograms per liter (ug/L). Milligrams per liter is a unit expressing the concentration of chemical constituents in solution as weight (milligrams) of solute per unit volume (liter) of water. One thousand micrograms per liter is equivalent to one milligram per liter. For concentrations less than 7,000 mg/L, the numerical value is the same as for concentrations in parts per million.

EVALUATION OF NONPOINT-SOURCE CONTAMINATION, WISCONSIN: SELECTED STREAMWATER-QUALITY DATA, LAND-USE AND BEST-MANAGEMENT PRACTICES INVENTORY, AND QUALITY ASSURANCE AND QUALITY CONTROL, WATER YEAR 199? By S.R. Corsi, J.F. Walker, D.J. Graczyk, S.R. Greb, D.W. Owens, and K.F. Rappold ABSTRACT The objective of the watershed-management evaluation monitoring program in Wisconsin is to evaluate the effectiveness of the best-management practices (BMFs) for rural streams, urban streams, and urban storm sewers. This report is an annual summary of the data collected for the program and a report of the results from several different special studies conducted within this program. Suspended sediment and total phosphorus storm-load data are summarized for eight rural sites and suspended sediment, total phosphorus, total recoverable lead, total recoverable copper, total recoverable zinc, and total recoverable cadmium storm-load data are summarized for four urban sites. Dissolved-oxygen data is summarized and compared with Wisconsin's waterquality standards for summer 1993 for seven rural sites. The dissolved-oxygen concentrations declined to levels below these standards at least one time at all seven sites during summer 1993. Total-recoverable hardness concentrations were compared with dissolved-hardness concentrations at two urban streams and two urban storm sewers. Least-squared linear regressions resulted in stronger relations for low-flow conditions than for high-flow conditions, indicating that most hardness during low flow is dissolved hardness. Pesticide data are summarized for four urban sites and six rural sites. Herbicides were detected at urban and rural sites; whereas insecticides were detected only at urban sites. A land-use and best-management-practice inventory is ongoing for each evaluation monitoring project to track the different sources of nonpoint pollution in each watershed and to document implementation of best-management programs that may cause changes in water quality of streams. Updated information is gathered each year, mapped, and stored in a geogrrohic-information-system data base. The quality-assurance/quality-control plan for the urban watershed-management evaluation program consisted of a series of blanV samples. These blank samples were used to identify and isolate contamination by inorganic and organic components throughout the collection and processing of urban streamwater samps. A dissolved trace-metal contamination pnblem was identified and resolved by using different laboratory-supplied sample bottles. A special study was done to determine the effect of holding time on fecal coliform colony counts. A linear regression indicated that the mean decrease in colony counts over 72 hours was 8.2 percent per day. Results after 24 hours showed that colony counts increased in some samples and decreased in others. INTRODUCTION In October, 1989, the U.S. Geological Survey (USGS) began a watershed-manageirent evaluation monitoring program in cooperation with the Wisconsin Department of Natural Resources (WDNR). The overall objective of each individual project in the program (fig. 1) is to determine if the water chemistry in the receiving stream has changed as a result of the implementation of land-management practices in the watershed. This is accomplished through monitoring of water chemistry and ancillary variables before best-management practices (BMPs) are implemented, during implementation, and after watershed-management plans have been completely implemented. The period before BMP implementation is termed "pre-BMP," the period during active implementation is termed "transitional," and the period after complete implementation is termed "post-BMP."

92' 46° 88° 87" 45° EXPLANATION T Rural watershed evaluationmonitoring site Urban watershed evaluationmonitoring site 100 MILES 100 KILOMETERS WISCONSIN Rattlesnake Creek and Kuenster Creek Base from U.S. Geological Survey National Atlas, 1970 "1 Lincoln Creek and Menomonee River Nine Springs Creek tributary s*orm sewer and Monroe Street detention pond Figure 1. Map showing location of rural and urban sites in the Wisconsin watershed-management evaluation monitoring program, water year 1993. The county Land Conservation Departments (LCD's) and the WDNR have identified sources of nonpoint pollution in each rural watershed (table 1). This information was used to select sites that are eligible for partial funding of BMP implementation. The LCD's are in the process of contacting land owners to request that they implement the appropriate BMFs for streamwater quality improvement. This is a voluntary program and, therefore, may produce variable success depending largely on the percentage of land owners that implement the recommended BMFs. The WDNR and each city have identified sources of nonpoint contamination in the urban watersheds (table 1). Nonpoint pollution reduction goals have been set, but a specific plan identifying the type and location of BMFs needed to achieve these goals has not been defined. This report, the second in a series of annual progress reports, is divided into three sections and includes four appendixes. The following topics are addressed: (1) streamwater-quality data, constituent loads in storm runoff, dissolved oxygen, hardness, and pesticides, (2) land-use and BMP inventory, and (3) quality assurance and quality control and effect of sample holding time on survival of fecal coliform colonies'. In each section, data collected during water yeir (WY) 1993 (October 1, 1992 through September 30, 1993) are presented and, if appropriate, implications for future data-collection efforts are discussed. The appendixes present the storm-load data collected during WY 1985-93 and the qualityassurance/quality-control document developed during WY 1993 for sampling at the urban sites. SELECTED STREAMWATTR- QUALITYDATA In this section, the streamwater-quality data collected during WY 1993 are summ?rized in four parts. The first part describes the estimated con-

Table 1. Location, site type, and principal function of sites in the watershed-management evaluation monitoring program, Wisconsin, water year1 1993 Site name Latitude Longitude Site type Principle site function Black Earth Creek at County Trunk P Black Earth Creek at Mills Street Black Earth Creek at South Valley Road Bower Creek Brewery Creek Eagle Creek Garfoot Creek Joos Valley Creek Kuenster Creek Lincoln Creek Menomonee River Monroe Street detention pond inlet Nine Springs Creek tributary storm sewer Otter Creek Rattlesnake Creek 43"06'38" 43°06'48" 43°07'30" 44°25'21" 43°07'09" 44°12'34" 43°06'37" 44°12'54" 42°47'27" 43°05'49" 43°02'44" 43°03'09M 43°02'03"

42°46'49" 89°38'44" 89°39'00" 89°42'35" 87°56'24" 89°38'25" 91'40'42" 89°40'46" 91°39'54" 90°57'26" 87°58'20" 87°59'59" 89°26'07" 89°23'35" 87°55'20" 90°56'32" rural rural rural rural rural rural rural rural rural urban urban urban urban rural rural dissolved oxygen dissolved oxygen dissolved oxygen water chemistry water chemistry water chemistry water chemistry and dissolved ogen water chemistry water chemistry and dissolved oxygen water chemistry water chemistry water chemistry water chemistry water chemistry and dissolved ogen water chemistry and dissolved oxygen xWater year is the 12-month period, October 1 through September 30. The water year is designated by the calendar year in which it ends. Thus, the year starting October 1, 1992 and ending September 30, 19P3 is called the "1993 water year." stituent loads and summarizes the efficiency of the sampling protocols. The second part summarizes dissolved-oxygen data. The third part presents the relation of sampling total-recoverable and dissolved-hardness data from the urban sites. The final part summarizes the pesticide data collected during WY 1993. Constituent Loads in Storm Runoff Streamwater quality was monitored at eight rural and four urban sites (fig. 1), at base flow and during storms. The water-quality data and discharge data were used to estimate total constituent loads for the storm periods. Eventually, the storm-load data will be used to evaluate the effect of BMFs on streamwater quality. In general, at least 20 pre-BMP and 20 post-BMP storms are needed to detect moderate differences related to implementation of the BMFs (Walker, 1993, 1994). Summary of Data The watershed-management evaluation monitoring sites currently have between one year

(Lincoln Creek) and 6 years (Brewery and Garfoot Creeks) of data. Precipitation, streamflow volume, and storm loads for several different water-quality constituents have been compiled (Appendixes 1 and 2). Summary of Storm-Sampling Efficiency Suspended-solids and total-phosphorus storm loads were calculated at the eight rural sites for all storms having a sufficient number of samples to accurately define the distribution of concentration over time. With the exception of two sites (Bower and Otter Creeks) ammonia-nitrogen loads also were calculated for all of the rural sites. The total number of storms sampled at the rural sites during WY 1993 ranged from 15 to 27, and the median was 17.5-a sufficient number of storms for the overall evaluation of the watershed-management evaluation-monitoring program. Although a sufficient number of storms were sampled during WY 1993, additional hydrologic records were analyzed to evaluate the effectiveness of sampling procedures. For each site, the continuous streamflow record was inspected to identify all periods where a hydrograph rise and subsequent fall was significant. The hydrologist charged with maintaining each site defined each period of significant rise and fall, herein called hydrograph-rise period. The hydrograph-rise period is defined to be between the two points on the hydrograph where the storm begins and baseflow is resumed. For the purposes of this study, the beginning of the storm is the point of the first significant increase in discharge, and baseflow is resumed after the hydrograph-rise period is at a point where the hydrograph slope approaches the slope of baseflow recession. For each hydrographrise period, total runoff was computed, and each period was classified into one of four categories: (1) complete sampling, (2) equipment malfunction, (3) partial sampling, or (4) no sampling. Some of the equipment malfunctions during hydrograph-rise periods resulted in partial sampling and some resulted in no sampling. All hydrograph-rise periods classified with equipment malfunctions had an insufficient number of samples to define the water quality. The distinction between complete and partial sampling was determined individually by site on basis of the shape of the hydrograph and the number of samples collected. Storm events and runoff sampling categories for each of eight rural sites are shown in figures 2-9. In general, storm-sampling protocols appear to provide a representative set of data, in terms of storm frequency and magnitude and the season of occurrence. A high percentage of large storms, which tend to carry the greatest loads, were sampled completely during WY 1993. Potential deficiencies in the sampling protocol identified previously (Graczyk and others, 1993) were modified at the beginning of WY 1993. As a result of the modifications, samples for all of the sites are more representative than those collected during previous years. One significant improvement was the adjustment cf sampling thresholds during the winter to catch midwinter snowmelt. This adjustment resulted in more hydrograph-rise periods being sairnled, thus greatly improving the data set available for evaluation of the nonpoint program. A potential deficiency is the exclusion of several midsized storms during May and June at most of the sites. This exclusion was the result of a sho-tfall in the budget for sample analysis; the decision was made jointly by the cooperator and project personnel to sample only the large storms in May and June until a new contract went into effect in July. Monitoring activities in Water Year 1994 Although it appears that an adequate number of hydrograph-rise periods are being sampled, the efficiency of sampling protocols will continue to be monitored closely. An analysis similar to that presented above will be done periodically so that sampling protocols can be fine-tuned when necessary. In an effort to continue effective sampling of the midwinter snowmeh, sampling thresholds will be monitored and adjusted as necessary. Dissolved Oxygen Dissolved-oxygen data were collected continuously at seven rural sites: Garfoot Creek, Black Earth Creek at County Trunk P, Flack Earth Creek at Mills Street, Black Earth Creek at South Valley Road, Otter Creek, Rattlesnake Creek, and Kuenster Creek (fig. 1). Dissolvedoxygen data were collected during open-water periods; all dissolved-oxygen meters were removed during the winter.

Illllllllllllllilllllllll en u 0.6 t-°'4 g 0.3 iTriiiriiiiiiiiiiiiiiiiiiii rrrii iITi i i T i i iiii ii i ii i COMPLETE m EQUIPMENT Q PARTIAL n NONE xl-Jlllninlln

111 i Xj.AMHj.J.J-J.lhihX IO Q C § T-CVJCOTth-OCpOCO BEGINNING OF PERIOD, MONTH/DAY h- h- s Figure 2. Graph showing total rainfall and runoff for selected hydrograph-rise periods at Brewery Creek monitoring site, Dane County, Wis., water year 1993. w£ 0.6 O 2 0.5 o.3 MMIirillllllllMlllMIIM ITT I I m I I fl] I I I I I I I I I I I A I I, i illi.illlnlJ..J.Il.li I I I I I I I I I I I I I I I I I I 1 I I I I I I I I I I I I I I I I I I COMPLETE SB EQUIPMENT U PARTIAL D NONE .nlnln mlnLl Iflnnnlli InJllLin T-T-T-OOOOOOOOOOOOOOOOOOOOOO BEGINNING OF PERIOD, MONTH/DAY Figure 3. Graph showing total rainfall and runoff for selected hydrograph-rise periods at Garfoot Creek monitoring site, Dane County, Wis., water year 1993.

. ll.ll Illllll .ll...lllll.l LJJ g 0.2 O I 0.1 O ri i i i i i i i COMPLETE EQUIPMENT H PARTIAL D NONE LnJ n n nJ BEGINNING OF PERIOD, MONTH/DAY Figure 4. Graph showing total rainfall and runoff for selected hydrograph-rise periods at Eagle Creek monitoring site, Buffalo County, Wis., water year 1993. Hffi LL O L , Ihi.lililh.il,..illll. uj 0.18 go.16 O2 0.08 O 0.02 COMPLETE gg EQUIPMENT Q PARTIAL inloi D NONE ME Jl M n nnnl OOOOOOOOO BEGINNING OF PERIOD, MONTH/DAY Figure 5. Graph showing total rainfall and runoff for selected hydrograph-rise periods at Joos Valley Creek monitoring site, Buffalo County, Wis., water year 1993.

O r r COMPLETE

EQUIPMENT Q PARTIAL ° 55588333S3388888 BEGINNING OF PERIOD, MONTH/DAY Figure 6. Graph showing total rainfall and runoff for selected hydrograph-rise periods at Bower Creek monitoring site, Brown County, Wis., water year 1993. tn£ 1.2 oz z aT 0.8 *0.4 r m COMPLETE SS EQUIPMENT Q PARTIAL Q NONE r n 1" 1'1 BEGINNING OF PERIOD, MONTH/DAY Figiire 7. Graph showing total rainfall and runoff for selected hydrograph-rise periods at Otter Creek monitoring site, Sheboygan County, Wis., water year 1993.

x 2 u. o

I.ill.i. llli.li.il W O £0.4 r i r i COMPLETE EQUIPMENT D PARTIAL Q NONE Flnl BEGINNING OF PERIOD, MONTH/DAY Figure 8. Graph showing total rainfall and runoff for selected hydrograph-rise periods at Kuenster Creek monitoring site, Grant County, Wis., water year 1993. :en ii 11 i i i i i i i i i i M M i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i ".ill* I I -ll.il..lMlllllllll.l..lll.h. Z 0.5 it 04 § 0.3 IT b 0.1 COMPLETE 8! EQUIPMENT D PARTIAL Q NONE I, i-T-OO OOOOOOOOOOOOOO BEGINNING OF PERIOD, MONTH/DAY Figure 9. Graph showing total rainfall and runoff for selected hydrograph-rise periods at Rattlesnake Creek monitoring site, Grant County, Wis., water year 1993.

Summary of Data Maximum, minimum, and mean concentrations of dissolved oxygen for each of seven sites for WY 1990-93 are listed in table 2. The maximum dissolved-oxygen concentrations during the 1993 water year ranged from 13.1 mg/L at Garfoot Creek to 16.5 mg/L at Otter Creek (table 2). The minimum dissolved-oxygen concentration ranged from 0.50 mg/L at Kuenster Creek to 5.2 mg/L at Black Earth Creek at South Valley Road. The mean dissolved-oxygen concentration in 1993 decreased from the mean in 1992 at all of the sites except for Rattlesnake Creek, Kuenster Creek, and Garfoot Creek. The mean dissolvedoxygen concentration for those three sites increased 0.7 mg/L at Rattlesnake Creek, 1.0 mg/L at Kuenster Creek, and 0.3 mg/L at Garfoot Creek. The State of Wisconsin's water-quality standards require a minimum dissolved-oxygen concentration of 5.0 mg/L for warmwater streams (maximum water temperature is greater than 24°C) and 6.0 mg/L for coldwater streams (maximum water temperature is less than 24°C) (Wisconsin Department of Natural Resources, 1992). The number of days concentrations were below these standards and the total number of days dissolved-oxygen concentrations were monitored during WY 1991-93 are listed in table 3. The duration that the dissolved-oxygen concentration is less than the State standards is also an important factor in the safety of aquatic organisms. A frequency analysis was done to determine the return period, in days, when the instantaneous dissolved-oxygen concentrations would be less than the State of Wisconsin standard for 1 hour. In summer 1993, the dissolvedoxygen concentration was 6.0 mg/L for 1 hour at Black Earth Creek at South Valley Road once every 75 days or about twice during the summer (May-September) (fig. 10). In comparison, during WY 1992, the dissolved-oxygen concentration was 6.0 mg/L for 1 hour once every 52 days or about three times during the summer (Graczyk and others, 1993). At Garfoot Creek in 1993, the dissolved-oxygen concentration was 6.0 mg/L for 1 hour once every 30 days on average, whereas in 1992, the dissolved-oxygen concentration was 6.0 mg/L once every 7 days on average (Graczyk and others, 1993). In summer 1993 at Rattlesnake Creek and Kuenster Creek, the dissolved-oxygen concentration was 5.0 mg/L for 1 hour once eve~y 155 days on average or once per summer (fig. 11). In 1992, the dissolved-oxygen concentration was 5.0 mg/L for 1 hour once every 5 days on average or about 30 times for the summer (Graczyk and others, 1993). The higher dissolved-oxygen concentrations indicate an improvement in water quality at these two sites, however, the krorovement may be a result of weather and flow conditions rather than implementation of BMPs. Monitoring activities in Water Yer 1994 Dissolved-oxygen monitoring is planned to continue at the warmwater sites to investigate whether severe dissolved-oxygen reductions monitored in past years can and do still occur. Also, continued monitoring may help determine if the improvement of dissolved-oxypen concentrations at Kuenster and Rattlesnake Creeks found in WY 1993 is due to weather and flow conditions or to implementation of BMPs. Total-Recoverable and Dissolved Hardness Stormflow and fixed-interval water-quality samples collected at the four urban sites have been analyzed for dissolved and total-recoverable hardness since November 1992. To-al-recoverable hardness is determined from a whole-water sample processed by means of a mild digestion, whereas dissolved hardness is analyzed from a filtered water sample. This comparison was done to determine whether dissolved hardness could be estimated from a linear relation with totalrecoverable hardness. Linear regression models were used to determine the relation between total-recoverable hardness and dissolved hardness for fixed-interval and stormflow samples. No relation between total-recoverable and dissolved hardness was found in stormflow samples collected at Nine Springs storm sewer, Monroe Street detention pond, and Lincoln Creek (fig. 12a and b), as indicated by the low coefficients of determination (R2). However, a relation was found for stormflow and fixed-interval samples from Menomonee River and fixed-interval samples from Lincoln Creek (fig. 12c and d). Therefore, total-recoverable hardnes? could potentially be used in the future to predict dissolved

Table 2. Summary of surface water dissolved-oxygen concentration data collected at Wisconsin watershed-management evaluation monitoring sites, water years 1990-93 [Concentrations are in milligrams per liter; --, no data] Water year Maximum Minimum Mean Garfoot Creek Black Earth Creek at County Trunk P Black Earth Creek at Mills Street Black Earth Creek at South Valley Road Maximum Minimum Mean Otter Creek Rattlesnake Creek Kuenster Creek hardness in stormflow and fixed-interval samples from the Menomonee River and in fixedinterval samples from Lincoln Creek. Linear regression results for stormflow samples collected at Nine Springs tributary storm sewer, Monroe Street detention pond, and Lincoln Creek show that the data are scattered, an indication that the solids in the stormwater contain magnesium and calcium in dissolved and particulate form (fig. 12a and b). Stormflow and fixed-interval samples collected at Menomonee River and fixed-interval samples collected at Lincoln Creek indicate that nearly all of the total-recoverable hardness is dissolved (fig. 12c and d). Stormflow samples from Menomonee River with high concentrations of solids do not contain a significant amount of particulate magnesium and calcium. Total hardness in stormflow and fixed-interval samples from the urban river sites-Lincoln Creek and Menomonee River-tend to be higher than that of samples from the storm-sewer sites-Nine Springs tributary storm sewer and Monroe Street detention pond (fig. 12). This tendency could be attributed to the elevated concentrations of hardness in ground-water contributions to the river sites whereas the sorm sewers are closed and there is minimal ground-water contribution. Fixed-interval samples have higher total hardness than stormflow samples. Stormwater runoff with lower total hardness is diluting the base flow at Lincoln Creek and Menomonee River (fig. 12b, c, and d). This finding is supported by the lower total hardness for the storm samples collected at Nine Spring tributary storm sewer and Monroe Street detention pond.

Table 3. Number of days that surface water dissolved-oxygen concentration was less than the State of Wisconsin standard at selected stream sites during water years 1991-93 Water year No. of days dissolved oxygen concentration was less than standard Coldwater Garfoot Total no. of days dissolved oxygen was monitored streams Creek Black Earth Creek at County Trunk P Black Earth Creek at Mills Street Black Earth Creek at South Valley Road No. of days dissolved oxygen concentration was less than standard Warmwater Total n->. of days dissolved oxygen was moHtored streams2 Otter Creek Rattlesnake Creek Kuenster

Creek oldwater streams typically have maximum stream-water temperatures less than 24.0°C. 2Warmwater streams typically have maximum stream-water temperatures greater than 24.0°C. Pesticides Nonfiltered water samples were analyzed for pesticides commonly used in the vicinity of six of the rural sites and four of the urban sites. Samples were collected during storms, and samples collected nearest the peak discharge were analyzed for pesticides. Automated refrigerated samplers equipped with Teflon Mined tubing and glass bottles were used to collect samples at Rattlesnake, Eagle, Otter, and Bower Creeks (rural sites) and at Lincoln Creek, Menomonee River, Monroe Street detention pond, and Nine Springs tributary storm sewer (urban sites). To collect samples from Garfoot and Brewery Creeks (rural 1Use of brand, firm, and trade names in this report is for descriptive purposes only and does not constitute endorsement by the U.S. Geological Survey. sites), investigators waded to near th? center of the creek and dipped the sample bottle just below the water surface. Samples were chilled and shipped to the Wisconsin State Laboratory of Hygiene (WSLH) where they were analyzed. Generally, herbicides are less toxic to fish and other aquatic life than are most irsecticides. Researchers have found increased herbicide concentrations in waterways and runoff from nonpoint sources after herbicide application and during spring and early summer rainstorms (Thurman and others, 1992). Summary of Data Rural sites The nonfiltered samples were analyzed for the pesticides listed in table 4. These pesticides are insecticides and herbicides that are actively

£ 12 LU Q. W 10 gg 2z o LU Black Earth Creek at South Valley Road Black Earth Creek at County Trunk P Black Earth Creek at Mills Street Garfoot Creek State of Wisconsin dissolved-oxygen-concentration standard LU O Q LU RETURN PERIOD, IN DAYS Figure 10. Return period in days that the dissolved-oxygen concentration was less than a given concentration for one continuous hour, summer 1993, at selected coldwater stream sites in Wisconsin. LU 11 tr 10 LU tr § 8

z 7 R 6 Z LU Q LU CO Q Kuenster Creek Otter Creek Rattlesnake Creek State of Wisconsin dissolved-oxygen-concentration standard RETURN PERIOD, IN DAYS Figure 11. Return period in days that the dissolved-oxygen concentration was less than e given concentration for one continuous hour, summer 1993, at selected warmwater stream sites in VTisconsin.

IT LU J tr UJ ws tr (D A - Stormflow D Nine Springs tributary storm sewer A Monroe Street detention pond -0.014x + 37 R*2 0.0017

D D

A D -ttffrV ; T B - Stormflow Lincoln Creek 0.34x + 39.1 RA2 0.17 100 150 200 250 300 350 400 450 500 100 150 200 250 300 350 450 500 W CO UJz Q UJ

Q C - Stormflow Menomonee River 1.19x -76.3 R*2 0.79 I / I D - Fixed-interval Samples V Menomonee River Lincoln Creek 1.00x -8.1 0.97 100 150 200 250 300 350 400 450 500 100 150 200 250 300 350 400 450 500 TOTAL RECOVERABLE HARDNESS, IN MILLIGRAMS PER LITER Figure 12. Graphs showing relation between total-recoverable hardness and dissolved hardness of stormflow and fixed-interval water samples from the urban watershed-management evaluation monitoring sites, Wisconsin, water years 1992 and 1993. being applied in the watersheds. They were chosen for analysis after consultations with county agents in each basin. Detectable concentrations of herbicides were found in the water-sediment samples collected at all of the sampling sites. The highest concentrations of cyanazine and metolachlor were found in Garfoot Creek at concentrations of 72 [ig/L and 57 ug/L, respectively (table 4 and fig. 13). The highest concentration of alachlor was found at Brewery Creek, at 32 ng/L. Concentrations of atrazine were highest in Eagle Creek and Bower Creek, at 22 ug/L and 13 ug/L, respectively (table 4 and fig. 13). At Otter Creek, concentrations of atrazine were found at or near the limit of detection of 0.10 ng/L (table 4 and fip. 13). Concentrations of atrazine at Garfoot Creek were

Table 4. Summary of pesticide data in nonfiltered surface water samples from the six rural watershed-management evaluation monitoring sites in Wisconsin, water year 1993 [Units are micrograms per liter; less than limit of detection] Pesticide Number of samples Maximum Minimum Eagle Creek 2,4-D1 Atrazine1 Alachlor1 Cyanazine1 Dicamba1 Metolachlor1 Pendimethalin1 Trifluralin1 Carbofuran2 Chlorpyrifos2 Cis-Permethrin2 Dimethoate2 Fonofos2 Methomyl2 Parathion2 Phorate2 Terbufos2 Trans-Permethrin2 Number of samples Maximum Minimum Garfoot Creek Number of samples Maximum Minimum Brewery Creek

Table 4. Summary of pesticide data in nonfiltered surface water samples from the six rural watershed-management evaluation monitoring sites in Wisconsin, water year 1993~Continued Pesticide Number of samples Maximum Minimum Rattlesnake Creek 2,4-D1 Atrazine1 Alachlor1 Cyanazine1 Dicamba1 Metolachlor1 Pendimethalin1 Trifluralin1 Carbofuran2 Chlorpyrifos2 Cis-Permethrin2 Dimethoate2 Fonofos2 Methomyl2 Parathion2 Phorate2 Terbufos2 Trans-Permethrin2 Number of samples Maximum Minimum Otter Creek Number of samples Maximum Minimum Bower Creek Ol *IIerbicide 2Insecticide

IUU tr HI A U.I

tr HI cos tr

A r : A ' A A A A it ' ' A A A A Atrazine ! A A A ]

O Cd Cd

CD tr £ £ £ £ o o o o S o o

S a o" Z t

(8 m CO O ° ,£ HI CO Q a g £ oI H ,0 HI0 0.01 - A A A A t

A A A . A : A A A A , ,

CD CD ' ' : Metolachlor Limit of Detection CD d) CD CD CD CD £ CD o o

£ c? a HI CO CD -2m CD CD o m tr A A

A CD CD Garfoot A A A A A A A CD O) (8 HI A A A A

A A A A Alachlor : A A A.

A Limit of Detection ' ttlesnake (8tr CD £o (D 1m (D CD

m 0£ o CDs Cyanazine

A A A A A A A A A A A Limit of Detection t CD£ Garfoot CD CD CD O) (8 HI "S ttlesnake (8tr (D£ o 1 £ m m£ o CDi A : -A "CD

CD § Figure 13. Concentrations of atrazine, alachlor, metolachlor and cyanazine in water samples at the six rural watershed-management evaluation monitoring sites, Wisconsin, water year 1993.

also low; maximum concentration was 0.86 ug/L. Atrazine concentrations at Brewery Creek were higher than concentrations at Garfoot Creek; maximum was 3.6 ug/L. (Garfoot Creek and Brewery Creek watersheds are adjacent to each other and both are tributary to Black Earth Creek. The land uses in both watersheds are similar; however, atrazine is banned in the Garfoot Creek watershed (Wisconsin Department of Agriculture, Trade, and Consumer Protection, written commun., 1993).) Samples from Otter Creek had relatively low concentrations of pesticides compared to the other rural sites (table 4). Concentrations may be low because Otter Creek flows through two lakes that may dilute the pesticide concentrations. Two other herbicides were detected in samples from the rural sites. Detectable concentrations of dicamba and 2,4-D were found at all of the sites with two exceptions. At Garfoot Creek, no dicamba was found above the limit of detection (0.20 ug/L) and at Otter Creek, 2,4-D was not found above the limit of detection (0.50 ug/L). At Brewery Creek, maximum concentrations of dicamba and 2,4-D were 5.8 ug/L and 3.9 ug/L, respectively. The highest concentration of dicamba was 17 ug/L, detected in a sample collected from Bower Creek. All samples contained insecticide concentrations below the analytical reporting limit at all of the sampling sites (table 4). Urban sites Nonfiltered samples were analyzed for the pesticides listed in table 5. These pesticides are insecticides and herbicides commonly used in urban areas. They were chosen for analysis after consultations with lawn-care companies, lawn and garden stores, and hardware stores. Herbicides were detected more frequently and at higher concentrations than insecticides. Atrazine, alachlor, and 2,4-D were detected above limits of detection at all four urban sites (fig. 14), cyanazine was detected at Nine Springs Creek tributary and the Menomonee River, and dicamba was detected at the Menomonee River. The maximum atrazine, 2,4-D, cyanazine, and dicamba concentrations were 0.42 ug/L, 1.8 ug/L, 1.7 ug/L, and 0.49 ug/L, respectively, at Menomonee River. The maximum alachlor concentration was 0.62 ug/L at the Monroe Street detention pond. The difference in magnitude of concentrations from site to site was not substantial, but concentrations of herbicides at the urban sites were noticeably less than concentrations of herbicides at the rural sites. This difference between urban sites and rural sites is most likely due to the heavier application of the herbicides in agricultural areas. At the Monroe Street detention pond, three insecticides (diazinon, lindane, and chlordane) were detected. Maximum concentrations of diazinon, lindane, and chlordane in samples collected at the four urban sites were 0.49, 0.018, and 0.25 ug/L, respectively, all in samp'es from the Monroe Street detention pond. At Nine Springs Creek tributary and Menomonee Rivr, P,P'-DDT was the only insecticide detected. Tl e maximum concentration of P.F-DDT was 0 05 ug/L at Menomonee River, which is between the limit of detection and the limit of quantification. P,P'-DDT was banned from general use on January 1, 1973. Its presence in one sample at the Menomonee River and one sample at Nine Springs Creek tributary indicates that P,F-DDT is still being used or that residuals from before 1973 are still present in some places. At Lincoln Creek, no insecticides were detected (table 5). Monitoring activities in Water Year 1994 At all rural sites, samples for herbicides are planned to be collected for anothe~ year. The number of analytes, however, probably will be reduced. Analyses for insecticides are likely to be discontinued because none of the insecticides were detected in these samples. Most of the samples were collected in early June and July, a time that coincided with crop planting and pesticide application. Herbicide concentrations were low (near the limit of detection) after a major storm in early July at all of the sites; therefore, sampling beyond early July probably will be d :continued, especially if several substantial storms have occurred before July. At urban sites, samples for pesticides are planned to be collected at all of the sites for another year. The suite of however, probably will be reconsidered. Pesticides that have not been found above the limit of detection may be discontinued unless the limits of detec17

Table 5. Summary of pesticide data in nonfiltered surface water samples from four urban watershed-management evaluation monitoring sites in Wisconsin, water year 1993 [Units in micrograms per liter; less than the limit of detection] Pesticide Number of samples Maxi- Minimum mum Monroe Street detention pond 2,4-D1 Alachlor1 Atrazine1 Captan1 Cyanazine1 Dacthal Dicamba (Mediben, Banvel Pendimethalin1 Trifluralin1 Chlordane2 Chlorpyrifos2 Diazinon2 Dimethoate2 Disulfoton2 Lindane2 Malathion2 Methoxychlor2 P,F DDT2 Sevin2 <.l <.l <.3 <.3 <.2 <.2 Number of samples Maxi- Minimum mum Nine Springs Creek tributary <.l <.l <.2 <.2 Number of Maxi- Minisamples mum mum Lincohi Creek <.l <.l O O O <.2 <.2

Number of samples Maximum Menomonee <.2 Minimum River <.l <.l <.3 <.2

ce 1Herbicide 2Pesticide

cfui 4t 2.5 CM"-J °m 2'° QSg 1 , Uj£E Z5 1.0 A ¥ o s. 'CD Q CD P A A 'Cto

Co A A A Cd Cd O

A

A

; sir CD CD g E O Q EXPLANATION A Concentration of pesticide V Concentration of pesticide is less than limit of detection -limits of detection may vary between samples CD EC OCE 5t <LU U.QOco

FRATIOI ROGRA zo iz O u./ 0:3 n ; A -A A A : : i J

i" A j t £ to Q § o

O) CO CD LLJzee HIT <UJ U.Q. Oco ZS HITsi o5 zz

o EC CD CD A A : A A

A

A A A A

Cd

CD CD § CO Figure 14. Concentrations of alachlor, atrazine and 2,4-D in water samples at the four urban watershed-management evaluation monitoring sites, Wisconsin, water year 1993.

tion are lowered. Samples were collected from May through August 1993. The data include some pesticide detections through August, when pesticide sampling was discontinued. Also, according to lawn-care companies, pesticides are applied on lawns from April through September. Therefore, in future years, an effort probably will be made to collect at least one event sample per month for pesticides from April through September. LAND USE AND BEST-MANAGE- MENT-PRACTICE INVENTORY The Priority Watershed Program is administered by the WDNR and the Wisconsin Department of Agriculture, Trade and Consumer Protection to improve streamwater quality in Wisconsin. Designation of a priority watershed is based on severity of pollution, sources of nonpoint pollution contributing to degradation of the water-body, cooperation of local governments to provide support in the planning and BMP implementation phase, and support of local communities. After priority watersheds are chosen and priority watershed plans are written (Wisconsin Department of Natural Resources, 1989, 1990, 1991, 1993; Wisconsin Department of Natural Resources, written commun., 1990), the county Land Conservation Departments (LCD's) follow several steps to implement these plans. First, the stream is inventoried to determine its condition and to determine what land uses currently exist in the basin. Next, the inventory is broken down into sites that are eligible for BMP implementation based on the contribution of contamination to the stream from each site. Then the LCD contacts the land owners of eligible sites for a voluntary sign-up. If the land owner volunteers for the program, cost share agreement and design of BMP s are made with contingency for the practice to be installed by a certain date. The WDNR and the USGS implemented the watershed-management evaluation monitoring program in 1984 to test if the BMFs change streamwater quality. Eight rural sites and four urban sites from the Priority Watershed Program were chosen as test sites. Each of the test sites is a smaller watershed that is monitored within one of the priority watersheds. Eight rural sites were chosen with land characteristics similar to the eight rural test sites for use as reference sites. BMFs will not be implemented in the reference sites because they are not in a priority watershed. These sites are to be used as a parallel comparison to the test sites to helf describe anomalies found in streamwater quality data from the test sites. Progress in BMP implementation, changes in land use, and other watershed characteristics are being tracked for each watershed throughout the course of water-quality sampling. Thi information, along with the results from waer-quality analyses, will help to determine the cause of changes in water quality and to what extent BMFs should be implemented in order to achieve specified levels of water-quality improvement. Data Collection Updates on eligible and implemented BMFs, priority-watershed-plan inventory data, and some information on other land ixses were obtained from the LCD's for the indhadual test sites within each priority watershed, fome additional land-use information was obtained from the appropriate county LCD's for the individual reference sites. Road and bridge construction, along with general maintenance information for tl 3 test and reference sites, was requested from the appropriate county highway commissions and the Wisconsin Department of Transportation. This information will be used to augment the prioritywatershed-plan inventory data on nonpoint-pollution sources. In preparation for a spring 1994 ephemeral gully inventory, county-soil-survey maps from the U.S. Soil Conservation Service (SCS) were used to locate possible sources of gully erosion and areas where the occurrence of ephemeral gullies are likely. U.S. Agriculture Stabilization and Conservation Service annual aerial photographs of the test sites were also used to facilitate the identification of ephemeral gullies. Agricultural Land Use Table 6 summarizes agricultural land uses listed in the priority-watershed-plan inventories for the watershed-management evaluation monitoring watersheds. Cover types for cropland are, by nature, subject to change; future uses of the

Table 6. Agricultural land use in selected rural watersheds in Wisconsin ["Other" may include residential, commercial, non-farm natural resources (such as lakes and woods), and any missing watershed data; --, cover type not found] Total acreage of cover type/percentage of total area Cover type Corn Hay Oats Corn (no rotation) Small grains Grassland Grazed woodlot Pasture Woodlot Wetland Farmstead Other Total Bower Creek (1988-89) 6,275/66 163/2 97/1 512/5 44/0 235/3 2,146/23 9,472/100 Otter Creek (1987) 1,420/24 1,062/18 1,404/24 578/10 11/0 748/13 345/6 154/3 121/2 5,843/100 Rattlesnake Creek (1989) 15,802/58 107/0 3,887/14 2,056/8 3,061/11 422/2 657/3 1,105/4 27,097/100 Efprle Creek (1988) 1,674/25 2,031/31 273/4 110/2 34/0 337/5 944/14

110/2 1,149/17 6,662/100 fields must be ascertained from the rotation codes listed for the individual fields (Wisconsin Department of Natural Resources, 1989, 1990, 1991, 1993; Wisconsin Department of Natural Resources, written commun., 1990). For example, for fields listed as corn, oats, or hay, corn may be grown 1-3 years, oats, 1 year, and hay, 1-4 years. Thus, the cover-type data for cropland account only for the year in which the inventory was completed. Development of Geographic- Information-System Data Base A geographic-information-system (GIS) data base is being developed for test and reference sites as follows: A. Base-map data are entered into the GIS data base by digitizing the mapped data directly from USGS 7.5-minute quadrangles. The base maps include the basin outline, the drainage system, the major roads, and the locations of stream and rain gages. B. Information on eligible and implemented BMPs, land uses, and other changing watershed characteristics is obtained from the county LCD or the SCS for each individual site. C. The eligible and implemented BMFs, land uses, and other changing watershed characteristics are then digitized and incorporated onto the base maps. D. The maps are updated each j'ear to incorporate changes in any of the mapped information. The eight rural reference sites were digitized and development of base maps was started. The two urban test sites and the eight rural test sites were digitized, and base maps wer? completed. Land-use and priority-watershed-plan inventory data were digitized and displayed on the Otter Creek base map. Activities in Water Year 194 All pertinent land-use data will be entered into a computer data base at the WDNR by a

USGS employee. An updated contaminant source evaluation of the test sites will be developed. This evaluation will account for changes that have occurred since the original priority-watershedplan inventory and will be accompli shed by use of data supplied by the LCD's, by completion of new field inventories of nonpoint-source contaminants (for example, gully and streambank erosion), and by generation of new nonpointsource-contaminant loads from computer models. Control and development of the GIS data base will be transferred from the USGS to the Bureau of Information Management (BIM) within the WDNR. The BIM is expected to assume the primary responsibility for future development of the GIS data base. QUALITY ASSURANCE AND QUALITY CONTROL The quality-assurance/quality-control (QA/ QC) plan for the rural watershed-management evaluation monitoring sites was defined previously (Graczyk and others, 1993), and the QA/QC plan for the urban watershed-management evaluation monitoring sites beginning in the spring of 1994 is outlined in Appendix 3. Before the spring of 1994, the QA/QC plan for the urban sites is not formally defined. The field procedures, sample processing, and equipment cleaning procedures during 1993 were the same as outlined in Appendix 3. Blank samples were treated somewhat differently. Before the spring of 1994, each project chief took several blank samples per year to determine if contamination existed in the samples. When contamination was found, blanks were taken for each individual component of the sampling process defined on page 54. Inorganic and Organic Constituents at Urban Sites The QA/QC plan for the urban watershedmanagement evaluation monitoring project during WY 1993 consisted of a series of blank samples collected from Milli-Q (analyte-free) water passed through different components used in sample collection and processing of streamwater samples from urban sites. These blank samples are used to attempt to isolate inorganic and organic contamination from five components: the sample bottle, the filtering apparatus, the automatic sampler, the filtering pump tube, and the splitter. Blank samples are analyzed for the same constituents as those analyzed for in the streamwater samples. Inorganic Constituents Sample bottles Only one bottle blank was collected during WY 1993. In general, all concentrations were below the limits of detection. The few exceptions were specific conductance (3 (iS/cm), alkalinity (3 mg/L), and dissolved zinc (31 (ig/L). The blank concentrations should be similar to the concentration of Milli-Q water, which generally has a specific conductance less than 2 (iS/cm and alkalinity less than 1 mg/L. The high concentration of dissolved zinc was found to be a result of contamination in the 60-mL sample bottles supplied by the WSLH, as described below. Filter-blank samples Most of the blank samples collected were filter blanks (n=21) because early results indicated that the filters may be a source of contamination. The four constituents that were found as contaminants in the filter blanks were dissolved cadmium (Cd), dissolved copper (Cu), dissolved zinc (Zn), and dissolved phosphorus (P) (fig. 15). Beginning in August 1993, dissolved metal samples were filtered into 250-mL metals bottles, whereas earlier samples were filtered into 60-mL sample bottles. This change in bottles dropped the degree of metal contamination noted in the filter blanks, with one exception. Dissolved cadmium was detected in two sample? after sample bottles were changed. (The laboratory limit of detection decreased during this sampling period, and the appearance of a decline in filter blanks' metal concentrations (fig. 15) sho-ild not be attributed to improved QA/QC procedures.) A discussion of the details of this contamination problem is contained in the following section on Significance of Sample Contamination Concentrations of dissolved phosphorus were at or above the limits of detection in 10 samples. In only 2 out of the 10 contaminated samples, however, were concentrations greater than the limits of quantification (Appendix 4). The source of this contamination is unknown.

DISSOLVED PHOSPHORUS, IN MILLIGRAMS PER LITER DISSOLVED CADMIUM, IN MICROGRAMS PER LITER 03 H 3 M (6 Ol Is c& P oQ" P o 3 O a B go w o' O GO

P 3g

CO O O O O O p O O O O O O O O O O

ro co -P* ui

§

-n

(O Co

CO O O O O O ro

en bo o

gg, ' "

".

ffl* o ;t

ro

o M -

B'i i-1 P COC6 cop "c? DISSOLVED ZINC, IN MICROGRAMS PER LITER oo-ossiiissi O . co (O CO

Co O 'O "S3 (D 2 Co O Dissolved Copper, In Micrograms Per Liter

Splitter-blank samples Significance of Sample Contamination Two blanks were collected by splitting Milli-Q water through the Teflon-lined churn splitter. Inorganic contamination problems in both samples were similar to those found for the bottle and filter blanks (for example, dissolved copper, dissolved zinc, alkalinity). The relative contribution of contamination from the churn splitter could not be determined because of the bottle contamination. Pump-tube blank samples Two pump-tube blanks were collected by pumping Milli-Q water through the filter pump tubing and collecting the sample directly into 60-mL bottles. Contamination by dissolved cadmium and zinc was found in both samples, presumably originating from the 60-mL bottles. Consequently, contamination from the pump tubing itself could not be determined. Sampler-blank samples A total of six sampler blanks were reported. Because sampler blanks are a measure of the cumulative contamination throughout the entire sampling process, the contamination observed in any of the component blanks discussed above would also be detected in the sampler blanks. In addition to the contaminants previously found in association with the 60-mL bottles, various additional contaminants (calcium, chloride, nitrate, sulfate, suspended solids, and total solids) were observed in the sampler blanks. This contamination appeared to be random and not associated with any particular site or time period. Organic Constituents One splitter blank and one sampler blank were collected for analysis of 37 organic constituents sampled at the urban sites, including a number of polycyclic aromatic hydrocarbons (PAH'S) not included in this report. Four PAH's fluoranthene, and pyrene) were found in the sampler blanks at concentrations greater than the limits of detection but less than the limits of quantification. Only fluoranthene and pyrene were detected in the splitter blank. The concentrations of these two constituents were similar in the two blanks: fluoranthene, 0.0093 and 0.0092 ug/L; and pyrene, 0.0084 and 0.0080 ug/L.) The 60-mL bottles were clearly a source of the systematic dissolved-metal contamination in samples collected before August 1993. Because of the variation in this contamination, there is no simple way to correct for this error. Dissolved zinc analyses before this date are unreliable but concentrations of other dissolved-metals in the blank samples were near the limit oc detection, therefore analyses before this date are not necessarily unreliable. Unfortunately, dissolved cadmium contamination was still observed after the bottle problem was corrected in Aupust; therefore, this problem still needs to be closely watched. Since dissolved cadmium is detected at levels near the limit of detection in stream samples, even low levels of contamination are significant. Even though half of the f Iter blanks indicated some degree of dissolved-hosphorus contamination, the extent of this con tamination relative to the concentration of dissolved-phosphorus in streamwater was generally low. Two exceptions are filter-blank samples tl at had dissolved-phosphorus concentrations of 0.006 mg/L. Use of additional bottle and filter blanks in the future may help to define the extent cf this phosphorous contamination. The source of the high alkalinity of the bottle blank (3 mg/L) is still unknown. This1 blank was simply Milli-Q water poured directly in the sample bottles and the contamination problem must lie either with the alkalinity bottles with the Milli-Q water, with the field technique, or with the lab analysis. Because only one sample was collected, it is difficult to determine how serious a problem this may be. Given the fret that the Menomonee River typically has alkalinity concentrations that range from 100 to 20°> mg/L, the low level of sample contamination is probably acceptable. Possible contamination from the churn splitter and the pump tubes could not be determined because of the few samples collected and the overshadowing bottle contamination. The source of the seemingly random contamination by constituents such as caldum, alkalinity, suspended solids, and chloride in the sampler blanks was impossible to determine; however, this contamination is probably due to residue particulate matter and streamwater in the

samplers. Contamination by these constituents is not a serious concern because the level of contamination is generally one or two orders of magnitude less than the instream concentrations. Although the organic contamination was between the limits of detection and quantification, findings indicate a systematic contamination somewhere in the sampling process. Most streamwater samples had concentrations that were considerably greater than the contamination levels found in the two blank samples, but a few streamwater samples had concentrations quite similar to the blank concentrations. The organic-concentration data available for streamwater to date can be used for interpretive purposes, but future QA/QC work would be needed to isolate the source of this contamination. Contamination by either inorganic or organic constituents during the collection and processing of water-quality samples for the 1993 urban sampling program are not significant. On the other hand, the frequency of blank sample collection, especially collection of bottle blanks, must be increased if tenuous interpretations based on one or two blanks are to be avoided. If sampling personnel closely follow a more regimented schedule for collecting sample blanks, then it may be possible to determine when and where the contamination is occurring. Because the QA/QC efforts are extremely important in verifying the integrity of the waterquality data, results from blanks need to be obtained from the WSLH as soon as possible, and the staff from WDNR and USGS need to work collectively to correct situations such as the dissolved-metal-bottle contamination problems when they occur. The current mechanism of the annual review of blank results is inadequate in responding to quality-control problems. Given the large investment of time and money that goes into this sampling program, periodic staff meetings (maybe every month during the sampling season) are warranted to review, discuss, and take the necessary corrective measures to prevent future losses of data. The QA/QC plan outlined in Appendix 3 is scheduled to be implemented in spring 1994. Fecal Coliform Bacteria The work described in this section of the report is a continuation of a study began in late fall 1993. The objective of this study was to determine the effect of holding time on survival of fecal coliform colonies because field personnel sometimes found it difficult to deliver samples to the WSLH within the required 24-hour holding time. In August and September 1993, 20 samples were collected by USGS personnel during runoff events at Brewery Creek and Garfoot Creek (fig. 1). All of these samples were received at the WSLH within 24 hours of collection time (WSLH recommended maximum holding tire). Samples were set up in triplicate at four different holding times: 0, 24, 48, and 72 hours, where 0 hours is the time when the sample was received at the laboratory. Sample bottles were refrigerated between setup times. Each plate count was obtained by setting up a three-to-fou~ serial dilution sequence and choosing the optimal plate (preferably 20-60 colonies per plate). The total number of analyses was 240 (20 samples x 4 holding times x 3 replicates). Ten percent of the plates were reported as "too numerous to count" and could not be used in the data analysis. All of these plates were from the 0-hour holding time. Fecal coliform counts ranged from 100,000 to 4,400,000 colonies per 100 mL of sample and were considerably higher than counts determined in November 1992 (table 7). Ir both years, all samples far exceeded the State recommended limit of 200 colonies per 100 mL for full-body contact in recreational waters (data are on file at the U.S. Geological Survey in Madison, Wis.). Illustrated in figure 16 are examples of colony counts in the 1994 data. As was the case in the previous year (Graczyk and others, 1993), large variation was seen within each replicate (fig. 16). The mean coefficient of variation withir a replicate was 14.5 percent and replicates differed by as much as 43 percent. This variation did not exhibit a consistent pattern with holding time. This inconsistency is seen in the Brewery Creek sample (fig. 16), where the variability and the mean concentration appeared to be decreasing, but on the third day, the concentration substantially increased. As was the case in November ard December 1992, fecal coliform counts generally decreased during the 4-day investigation. Plotted in fig25

Table 7. Fecal coliform summary statistics, Brewery and Garfoot Creeks, Wisconsin, water years 1993-94 [Concentrations are in colonies per 100 milliliters] Site Brewery Creek Garfoot Creek Dates '92/11-92/12 93/8-93/9 92/11-92/12 93/8-93/9 Observations Mean 41,000 780,000 130,000 640,000 Minimum 8,300 110,000 28,000 100,000 Maximum 160,000 2,800,000 310,000 4,400,000 Median 25,000 615,000 97,000 420,000 Samples were collected in November and December, 1992 and August and September, 1993. ures 17 and 18 are the replicate means against holding times for all samples. Linear regression models of the log-concentration values (dependent variable) with respect to time (independent variable) were calculated for all curves. Negative slopes were found for all 20 samples. The mean slope (rate of coliform die off) was -8.2 percent per day. The mean slope for samples from November and December 1992 was slightly higher, -12.6 percent. Data Analysis In an attempt to explain the inconsistent patterns observed in means and variances such as those seen at Brewery Creek (fig. I8a), the counts and dilution factors were obtained from the laboratory and examined (table 8). The serial dilution used during this 3-day time series was not consistent (as mentioned previously, a number of plates of differing dilutions were set up for each replicate, and the optimum plates were chosen). The dilution factor for the Brewery Creek sample was initially 10,000:1 and was changed to 1,000:1 for one of the replicates after 24 hours, and all of the replicates after 48 hours. After 72 hours, a dilution factor of 10,000:1 was used for all three replicates. Increasing the dilution factor on the fourth day explains the increased variability seen on the fourth day. In this case, the change to the dilution factor also resulted in an apparent increase in the mean fecal coliform concentration, but this increase is actually an anomaly due to the large variability on the fourth day. Application of t-test A t-test was applied to the data to determine whether significant differences in fecal coliform counts exist with respect to hold'ig times. Because the t-test compares only two treatments, it was run three times, 0-against 24-hour holding time, 0-against 48-hour holding time, and 0-against 72-hour holding time. The level of significance was set at p 0.10, anc1 variances were assumed to be unequal. Differences in fecal coliform counts were significant for 3? percent of the samples concentration at 24 hours, 45 percent at 48 hours, and 50 percent at 72 hours. In the 0-against 24-hour holding time runs, two increases and two decreases in concentrations were significant. All of the significant differences in comparisons of 0 against 48 hours and 0 against 72 hours were decreases. The equal number of increases and decreases in the 0-against 24-hour comparison indicates that fecal coliform concentrations may have increased in the bottle. Gordon and Fliermans (1978) have demonstrated that fecal coliform colonies can grow in the aquatic environment, outside of the host animal. The samples from Brewery and Garfoot Creeks may be especially susceptible to fecal coliform growth given the relatively high concentrations of nutrients found in these samples. Clearly, the longer holding period of 48 and 72 hours resulted in a fecal coliform die-off in the sample bottles. Unfortunately, contusions regarding the 24-hour additional holding time results are somewhat unclear, and most sample rejections due to holding-time problems are from samples that were held for only one additional day, not two or three. The 24-hour results indicate that either fecal coliform can die off or they can grow during this period, especifHy during warm weather. Therefore, in order to obtain

3,000,000 2,000,000 1,000,000 5 W cru- Stw t 500,000 tro: UJUJ mo. 200,000 100,000 Samples collected on August 15,1993 HOLDING TIME, IN DAYS 3,000,000 2,000,000 1,000,000 cruten -JOT Oui

500,000 S3 crcr tutu mo_ 200,000 100,000 Samples collected on September 14,1993 HOLDING TIME, IN DAYS Figure 16. Mean fecal coliform colony counts for different holding times of replicate water samples from Brewery Creek, Wisconsin, for two storms in 1993. Each line represents a different sample within the same storm.

w 10,000,000 LLJ ocu. ten -JOE OLD OH 1,000,000 i Si tree mui mo. 100,000 Samples collected on August 15, 1993 w 1,000,000 QS

tru.

8 m H Oil tro: LULU mo. 700,000 500,000 400,000 300,000 200,000 100,000 Samples collected on August 23, 1993 LU 3,000,000 2,000,000 g|±j 1,000,000 _,5 500,000 £§ mg; 200,000 100,000 Samples collected on September 14,1?93 HOLDING TIME, IN DAYS Figure 17. Mean fecal colifonn colony counts for different holding times of replicate water samples from Garfoot Creek, Wisconsin, for three storms in 1993. Each line represents a different sample within the same storm.

280,000 260,000 240,000 - 220,000 - h-UJ o 1-' jcn 200,000 - Otr 180,000 - §3 go 160,000 - 140,000 - 120,000 - 100,000 Sample collected on September 13,1993, at 2215 hours Day 0 Day 1 Day 2 HOLDING TIME, IN DAYS MAXI MUMMEAN MINIMUM - Day 3

tr=! Oo Oo UJQ. U. 500,000 450,000 400,000 350,000 300,000 250,000 200,000 150,000 100,000 B Sample collected on August 23, 1993, at 1700 hours Day 0 Day 1 Day 2 HOLDING TIME, IN DAYS fMlXIMUM MEAN 4MINUMUM Day 3 Figure 18. Examples of the effect of holding time on maximum, mean and minimum fecal coliform colony counts in water samples from (A) Brewery Creek and (B) Garfoot Creek, Wisconsin.

Table 8. Fecal coliform counts in, and dilutions used for, the water sample collected from Brewery Creek, Wisconsin, on September 13, 1994 [Concentrations are in colonies per 100 milliliters] Holding times 0 hours 24 hours 48 hours 72 hours Replicate Count Dilution Count Dilution Count Dilution 10,000 1,000 10,000 10,000 10,000 1,000 1,000 1,000 1,000 10,000 10,000 10,000 accurate fecal coliform data, it may be even more important to get the samples to the laboratory within the first 24-hour period and avoid not only attrition of the fecal coliform sample but potential growth and increase in fecal coliform concentrations. Conversely, for the small sample size tested here the numbers of increases and decreases were equal, and it is possible that these opposing factors will somewhat cancel each other out when pooled together over the course of the monitoring program. In addition, rejection of 24-hour-old samples may result in the substantial decrease in the number of samples, which will decrease the statistical power for detecting changes in long-term fecal coliform concentrations. Variance-component analysis The variability of fecal coliform concentrations due to the effect of holding time was examined in the context of the other sources of variability, namely the analytical (determined from the replicates) and sampling (time within a storm that samples were collected) variability, by use of a variance component analysis on the data set. The holding time and analytical variabilities are possible sources of error, whereas the sampling-time variability represents the natural variability of fecal coliform counts at different times during an event. The variability due to all three factors is listed in table 9. In all but one case, the variability of fecal coliform concentrations resulting from time of sampling within a given storm was the largest source of variability. In no instance was holding time the largest source of variability. Correlation of fecal-coliform concentrations with other water-quality variables Concentrations of fecal-coliform and other water-quality characteristics that were measured at the time of the fecal-colifom sampling were correlated to determine which factors are associated with fecal coliform concentrations. The Pearson correlation coefficients are listed in table 10. Water temperature had the highest correlation with fecal coliform concentrations. The fact that water temperature has a strong effect on fecal coliform has been observed by others (Gordon and Fliermans, 1978; Hunter and McDonald, 1991; Hirotani and Matsui, 1992). The significant correlations with total phosphor is, suspended solids, and volatile suspended sol ;ds suggest that fecal coliform concentrations may be associated with runoff or resuspension materials. Possible Methods to Improve Ana'ytical Results Statistical differences were observed in fecal coliform concentrations when sample bottles were held longer than the required maximum holding time (24 hours). These differences were apparent despite large variability anong replicates. Use of different dilutions for samples within the same time series resulted in misleading anomalies in fecal coliform concentrations over time. Use of consistent dilutions is warranted for any future time-series testing. These data provide some evidence of fecal coliform growth in the sample bottles at least during the first 24 hours (WSLH recommended maximum holding time). In order to obtain as rrany observations as possible for future statistical analyses, laboratory personnel could analyz? samples

Table 9. Values of the three major components of variance in fecal coliform concentre tions in water samples from Brewery and Garfoot Creeks, Wisconsin, water year1 1993 [Variances in billions of colonies squared] Site Brewery Creek: Garfoot Creek: and date August 15 September 14 August 15 August 23 September 14 Sampling 1,000 Source of variability time Holding time Analytical ear is the 12-month period, October 1 through September 30. The water year is de?ignated by the calendar year in which it ends. Thus, the year starting October 1, 1992 and ending Septembe~ 30, 1993 is called the "1993 water year." Table 10. Pearson correlation coefficients between fecal coliform concentrations in, and selected characteristics of, water samples from Brewery and Garfoot Creeks, Wisconsin, water year 1993 significant at 0.05 level] Variable Water temperature Biochemical oxygen demand (BOD) Ammonia (NH4) Total phosphorus Suspended solids Volatile suspended solids Correlation coefficient 0.76* .50* .46* .47* between 24 and 48 hours old and flag the results. Samples that have been delayed more than 48 hours could be discarded. SUMMARY suspended sediment, total phosphorus, total recoverable lead, total recoverable copper, total recoverable zinc, and total recoverable cadmium were computed for the urban sites. All storm load data are summarized in tables. The objective of the watershed-management evaluation monitoring program in Wisconsin is to evaluate the effectiveness of BMFs for rural streams, urban streams, and urban storm sewers. This report is an annual summary of the data collected for the program and a report of the results from several different special studies conducted within this program. Water-quality data from eight rural sites and four urban sites are summarized. Storm loads for suspended sediment and total phosphorus were computed for the rural sites, and storm loads for Continuous dissolved-oxygen data were collected during the summer of 1993 at seven rural sites. Resulting data are summarized in tables. Recurrence intervals are plotted and compared with Wisconsin's water-quality standards. The dissolved-oxygen concentrations declined to levels below these standards at least one time at all seven sites during WY 1993. Total-recoverable-hardness concentrations were compared with dissolved-hardn°ss concentrations at two urban streams (Lincoln Creek and Menomonee River) and two urban storm

sewers (Nine Springs Creek tributary storm sewer and Monroe Street detention pond inlet). Least-squared linear regressions were computed for samples collected during storm-flow and low flow. The storm sample regressions for the storm sewers and Lincoln Creek indicate very weak linear relations (coefficients of determination were less than 0.2) whereas the storm-sample regression for the Menomonee River indicates a fairly strong relation (coefficient of determination was 0.79). The low-flow sample regression for Lincoln Creek and Menomonee River was relatively strong (coefficient of determination was 0.97). The regression equation indicates that most of the hardness during low flow is dissolved hardness. Pesticide data were collected at four urban sites and six rural sites during WY 1993. Some herbicides were detected at the rural sites, but insecticides were not found above analytical reporting limits. Although herbicides were more prevalent, both herbicides and insecticides were found at the urban sites. Maximum and minimum concentrations of all pesticides are summarized in tables. A land-use and best-management-practice inventory is ongoing for each evaluation monitoring project to track the different sources of nonpoint pollution in each watershed. Information is being gathered from the county Land Conservation Departments, the county highway commissions, the Wisconsin Department of Transportation, and the U.S. Soil Conservation Service. This information is mapped and stored in a geographic-information-system data base. Each year the information for each watershed is reviewed and updated. The quality-assurance/quality-control plan for the urban nonpoint monitoring project during WY 1993 consisted of a series of blank samples collected from Milli-Q water passed through different components used in the sample collection and processing of urban stream samples. These blank samples were used to isolate inorganic and organic contamination from the different components of the sampling routine. Constituent concentrations in these blank samples indicated that some dissolved metal contamination in the early part of the water year was corrected during the later months of WY 1993. Some low level organic contamination was found. A special study was done to determine the effect of holding time on fecal-colifcrm colony counts. Samples were analyzed at 0 hours, 24 hours, 48 hours, and 72 hours. By use of the t-test and a level of significance of 0.10, a significant decrease in concentrations war found in 33 percent of the samples at 24 hours, 45 percent at 48 hours, and 50 percent at 72 hours. A linear regression shows the mean slope of concentration from 0 hours to 72 hours to be -8.2 percent per day. REFERENCES CITED Gordon, R.W., and Friermans, C.B., 1978, Survival and viability of Escherichia Coli in a thermally altered reservoir: Water Research, v. 12, p. 343-352. Graczyk, D.J., Walker, J.F., Greb, S.R., Corsi, S.R., and Owens, D.W., 1993, Evaluation of nonpoint-source contamination, Visconsin selected data for 1992 water year: U.S. Geological Survey Open-File Report 93-630, 48 p. Hirotani, H., and Matsui, Y., 1992, Positive correlations between catchment areas and densities of bacteria in the upper reaches of a river: Water Science and Technology, v. 26, no. 7, p. 1965-1972. Hunter, C., and McDonald, A., 1991, Seasonal changes in the sanitary bacterial quality of waters draining a small upland catchment in the Yorkshire Dales: Water Research, v. 25, no. 4, p. 447-453. Thurman, E.M., Goolsby, D.A., Mever, M.T., Mills, M.S., Pomes, M.L., and Kolpin, D.W., 1992, A reconnaissance study of herbicides and their metabolites in surface water of the midwestern United States using-immunoassay and gas chromatography/mass spectrometry: Environmental Science and Technology, v. 26, p. 2440-2447. Walker, J.F., 1993, Techniques for detecting effects of urban and rural land-us? practices on stream-water chemistry ir selected watersheds in Texas, Minnesota, and Illinois: U.S. Geological Survey Open-File Report 93-130, 16 p. Walker, J.F., 1994, Statistical techniques for assessing water-quality effects of BMPs: Journal of Irrigation and Drainage Engineering, v. 120, no. 2, p. 334-347.

Wisconsin Administrative Code, 1992, Rules of Department of Natural Resources environmental protection: Wisconsin Administrative Code S.NR. 100.01. Wisconsin Department of Natural Resources, 1989, A plan for the control of nonpoint sources and related resource management in the Black Earth Creek priority watershed: Wisconsin Department of Natural Resources WR-218-89, [variously paginated]. Wisconsin Department of Natural Resources, 1991, A plan for the control of nonpoint sources and related resource management in the Lower Grant River priority watershed: Wisconsin Department of Natural Resources WR-293-91, [variously paginated]. Wisconsin Department of Natural Resources, 1990, A nonpoint source control pirn for the Waumandee Creek priority watershed project: Wisconsin Department of Natural Resources WR-274-90, 150 p. Wisconsin Department of Natural Resources, 1993, A nonpoint source control pirn for the East River priority watershed project: Wisconsin Department of Natural Resources WR-274-93, 188 p.

Appendixes 1-4

Appendix 1. Storm-load data for rural watershed-management evaluation monitoring sites, Wisconsin [yr, year; mo, month; d, day; h, hour; in., inches; Mft3, million cubic feet; Ib, pounds; s/m, snowmelt; - water years1 1985-93 -, no data] Start of storm Date (yr/mo/d) 84/10/18 84/11/01 84/12/28 85/02/21 85/07/24 85/08/12 85/08/25 85/09/04 85/09/09 85/10/12 85/10/23 85/10/31 85/11/17 86/03/09 86/03/17 86/05/15 86/05/17 86/06/22 89/10/05 90/03/08 90/03/11 90/03/13 90/06/02 90/06/28 91/04/12 91/04/14 91/04/28 91/05/05 91/07/01 91/07/07 91/08/08 91/10/24 91/11/01 91/11/29 92/02/27 92/02/28 92/03/01 92/07/13 Time (24 h) End of storm Date (yr/mo/d) 84/10/19 84/11/01 84/12/28 85/02/25 85/07/26 85/08/13 85/08/26 85/09/05 85/09/09 85/10/13 85/10/24 85/11/02 85/11/19 86/03/10 86/03/20 86/05/16 86/05/18 86/06/22 89/10/05 90/03/09 90/03/12 90/03/14 90/06/03 90/06/29 91/04/13 91/04/14 91/04/29 91/05/05 91/07/02 91/07/08 91/08/08 91/10/26 91/11/02 91/11/30 92/02/28 92/02/29 92/03/02 92/07/15 Time (24 h) Brewery Precipitation (in.) Creek s/m s/m s/m s/m

s/m s/m s/m Streamflow volume (Mft3) Loads Suspendedsolids load (tons)

Appendix 1. Storm-load data for rural watershed-management evaluation monitoring sites, Wisconsin, water years 1985-93-Continued Start of storm Date (yr/mo/d) Time (24 h) End of storm Date (yr/mo/d) Time (24 h) Precipitation (in.) Streamflow volume (Mft3) Leads Suspendedsolids load (tons) Totalphosphorus load (Ib) Brewery Creek-Continued 92/08/29 92/09/16 92/09/18 92/10/15 92/11/20 93/03/06 93/03/07 93/03/08 93/03/16 93/03/24 93/03/25 93/03/26 93/03/27 93/03/28 93/03/31 93/06/07 93/06/17 93/07/05 93/07/07 93/07/09 93/07/17 93/07/25 93/07/27 93/08/15 93/09/13 92/08/29 92/09/17 92/09/19 92/10/15 92/11/22 93/03/07 93/03/08 93/03/09 93/03/17 93/03/25 93/03/26 93/03/27 93/03/28 93/03/29 93/04/01 93/06/08 93/06/18 93/07/07 93/07/08 93/07/10 93/07/18 93/07/26 93/07/28 93/08/16 93/09/15 s/m s/m s/m s/m s/m s/m s/m s/m s/m s/m 1,300 ' 4,000 Garfoot Creek 84/10/18 84/10/31 84/12/27 85/02/21 85/07/24 85/09/04 85/09/09 85/09/23 85/10/11 85/10/23 85/10/31 85/11/18 84/10/19 84/11/01 84/12/29 85/02/25 85/07/26 85/09/05 85/09/09 85/09/24 85/10/12 85/10/24 85/11/02 85/11/19 s/m s/m

Appendix 1. Storm-load data for rural watershed-management evaluation monitoring sites, Wisconsin, water years1 1985-93-Continued Start of storm Date (yr/mo/d) Time (24 h) End of storm Date (yr/mo/d) Time (24 h) Precipitation (in.) Streamflow volume (Mft3) Loads Suspendedsolids load (tons) Totalphosphorus load (Ib) Garfoot Creek Continued 86/03/09 86/03/16 86/05/15 86/05/17 89/10/05 90/01/16 90/03/11 90/03/13 90/03/14 90/06/02 90/06/28 90/08/19 91/03/01 91/03/22 91/04/12 91/04/14 91/08/08 91/11/01 91/11/29 92/02/26 92/02/27 92/02/28 92/09/16 92/09/18 92/11/19 92/12/15 93/03/24 93/03/25 93/03/26 93/03/27 93/03/28 93/03/31 93/04/07 93/04/15 93/04/19 93/06/07 93/06/17 93/07/05 86/03/11 86/03/20 86/05/16 86/05/18 89/10/06 90/01/17 90/03/12 90/03/14 90/03/15 90/06/03 90/06/29 90/08/20 91/03/02 91/03/23 91/04/13 91/04/14 91/08/08 91/11/02 91/11/30 92/02/27 92/02/28 92/02/28 92/09/17 92/09/18 92/11/21 92/12/16 93/03/25 93/03/26 93/03/27 93/03/28 93/03/29 93/04/01 93/04/08 93/04/16 93/04/20 93/06/08 93/06/18 93/07/07 s/m s/m s/m s/m s/m s/m s/m s/m s/m s/m s/m s/m

Appendix 1. Storm-load data for rural watershed-management evaluation monitoring sites, Wisconsin, water years1 1985-93-Continued Start of storm Date (yr/mo/d) Time (24 h) End of storm Date (yr/mo/d) Tune (24 h) Precipitation (in.) Streamflow volume (Mft3) Loads Suspendedsolids load (tons) Totalphosphorus load (Ib) Garfoot Creek-Continued 93/07/08 93/07/10 93/07/25 93/07/27 93/08/15 93/08/23 93/09/13 93/07/10 93/07/11 93/07/25 93/07/28 93/08/16 93/08/24 93/09/15 Eagle Creek 91/04/29 91/05/05 91/05/15 91/05/31 91/07/21 91/08/07 91/10/23 91/11/01 91/11/17 92/03/01 92/03/03 92/03/09 92/04/20 92/05/16 92/05/21 92/05/22 92/07/13 92/08/01 92/09/16 93/03/26 93/03/29 93/03/30 93/04/11 93/04/18 93/04/27 93/06/08 93/07/02 93/07/03 93/07/27 93/08/09 91/04/29 91/05/05 91/05/17 91/05/31 91/07/21 91/08/08 91/10/24 91/11/01 91/11/18 92/03/02 92/03/04 92/03/09 92/04/21 92/05/16 92/05/21 92/05/23 92/07/13 92/08/02 92/09/16 93/03/27 93/03/29 93/03/31 93/04/12 93/04/20 93/04/27 93/06/09 93/07/02 93/07/04 93/07/28 93/08/09 s/m s/m s/m 2,100 3,200 1,700 1,500 3,800 4,700 1,400 3,300 2,000 1,100 3,000

Appendix 1. Storm-load data for rural watershed-management evaluation monitoring sites, Wisconsin, water years1 1985-93-Continued Start of storm Date (yr/mo/d) 93/08/15 93/08/18 93/08/30 93/09/13 Time (24 h) End of storm Date (yr/mo/d) 93/08/15 93/08/18 93/08/30 93/09/14 Time (24 h) Eagle Precipitation (in.) Creek Continued Streamflow volume (Mft3) Lords Suspendedsolids load (tons) Totalphosphorus load (Ib) 1,000 Joos Valley Creek 90/08/17 90/08/26 91/04/29 91/05/05 91/05/15 91/05/31 91/07/21 91/08/07 91/10/23 91/10/31 91/11/17 92/03/01 92/03/03 92/03/08 92/04/20 92/05/16 92/05/21 92/05/22 92/06/17 92/07/02 92/07/13 92/07/22 92/08/01 92/09/16 93/03/24 93/03/26 93/03/30 93/04/11 93/04/16 93/04/18 93/04/27 93/06/08 90/08/18 90/08/26 91/04/29 91/05/05 91/05/17 91/05/31 91/07/22 91/08/08 91/10/24 91/11/01 91/11/18 92/03/02 92/03/04 92/03/09 92/04/21 92/05/16 92/05/21 92/05/23 92/06/17 92/07/02 92/07/14 92/07/23 92/08/02 92/09/16 93/03/25 93/03/27 93/03/31 93/04/12 93/04/17 93/04/20 93/04/28 93/06/08 s/m s/m s/m 1,600 1,500 1,700

Appendix 1. Storm-load data for rural watershed-management evaluation monitoring sites, Wisconsin, water years1 1985-93-Continued Start of storm Date (yr/mo/d) Time (24 h) End of storm Date (yr/mo/d) Time (24 h) Precipitation (In.) Streamflow volume (Mft3) Loads Suspendedsolids load (tons) Totalphosphorus load (Ib) Joos Valley Creek-Continued 93/07/01 93/07/03 93/07/27 93/08/09 93/08/15 93/08/18 93/08/30 93/07/02 93/07/03 93/07/27 93/08/09 93/08/15 93/08/18 93/08/30 1,600 Bower Creek 90/10/17 91/03/01 91/03/05 91/03/18 91/04/09 91/04/12 91/06/14 91/10/29 91/11/01 91/11/18 91/11/29 91/12/12 92/03/29 92/03/31 92/04/10 92/04/15 92/04/19 92/04/20 92/07/13 92/09/16 92/09/18 92/09/26 92/11/01 92/11/08 92/11/12 92/11/20 92/12/15 92/12/29 93/03/02 90/10/20 91/03/04 91/03/08 91/03/25 91/04/12 91/04/17 91/06/15 91/10/31 91/11/04 91/11/20 91/12/02 91/12/14 92/03/31 92/04/02 92/04/13 92/04/18 92/04/20 92/04/22 92/07/16 92/09/17 92/09/20 92/09/30 92/11/03 92/11/12 92/11/14 92/11/22 92/12/17 92/12/31 93/03/18 s/m s/m s/m s/m s/m 1,800 1,500 2,900 2,500 1,900 1,600

Appendix 1. Storm-load data for rural watershed-management evaluation monitoring sites, Wisconsin, water years 1985-93-Continued Start of storm Date (yr/mo/d) Time (24 h) End of storm Date (yr/mo/d) Time (24 h) Precipitation (in.) Streamflow volume (Mft3) Lozds Suspendedsolids load (tons) Totalphosphorus load (Ib) Bower Creek-Continued 93/03/24 93/04/04 93/04/07 93/04/11 93/04/15 93/04/19 93/04/27 93/05/30 93/06/08 93/06/14 93/08/05 93/03/31 93/04/07 93/04/10 93/04/14 93/04/19 93/04/23 93/04/29 93/06/03 93/06/11 93/06/16 93/08/08 s/m s/m s/m 1,100 1,100 1,700 2,400 1,300 3,600

Otter Creek 90/09/06 90/09/14 90/11/05 91/02/03 91/03/01 91/06/14 91/10/24 91/10/26 91/10/29 91/11/01 91/11/14 91/11/18 91/11/29 92/02/27 92/03/01 92/03/05 92/03/09 92/03/24 92/04/10 92/04/16 92/09/14 92/09/16 92/09/18 92/09/26 92/11/01 92/11/12 90/09/08 90/09/18 90/11/06 91/02/08 91/03/04 91/06/18 91/10/26 91/10/27 91/10/30 91/11/03 91/11/17 91/11/20 91/12/02 92/03/01 92/03/03 92/03/08 92/03/10 92/03/28 92/04/13 92/04/18 92/09/15 92/09/17 92/09/19 92/09/28 92/11/03 92/11/13 s/m s/m s/m s/m s/m

Appendix 1. Storm-load data for rural watershed-management evaluation monitoring sites, Wisconsin, water year?1 1985-93-Continued Start of storm Date (yr/mo/d) Time (24 h) End of storm Date (yr/mo/d) Time (24 h) Precipitation (in.) Streamflow volume (Mft3) Loads Suspendedsolids load (tons) Totalphosphorus load (Ib) Otter Creek-Continued 92/11/19 92/11/25 92/12/15 92/12/29 93/03/24 93/04/03 93/04/08 93/04/11 93/04/15 93/04/19 93/06/07 93/07/05 93/07/09 93/09/13 93/09/20 92/11/22 92/11/27 92/12/17 92/12/30 93/04/01 93/04/06 93/04/09 93/04/13 93/04/17 93/04/21 93/06/09 93/07/07 93/07/10 93/09/16 93/09/22 s/m s/m Kuenster Creek 92/11/01 92/11/19 92/12/15 93/03/05 93/03/07 93/03/08 93/03/15 93/03/25 93/03/26 93/03/27 93/03/28 93/03/30 93/05/02 93/06/17 93/06/28 93/06/29 93/07/17 93/08/14 93/08/18 92/11/04 92/11/22 92/12/17 93/03/06 93/03/08 93/03/09 93/03/17 93/03/26 93/03/27 93/03/28 93/03/29 93/03/31 93/05/03 93/06/18 93/06/29 93/06/30 93/07/17 93/08/16 93/08/19 s/m s/m s/m s/m s/m s/m s/m s/m s/m 1,100 3,000 1,200 3,600 1,400 1,190 6,700 1,400 1,600 Rattlesnake Creek 90/01/16 90/03/08 90/01/18 90/03/08 s/m s/m 1,300 1,600 3,600

Appendix 1. Storm-load data for rural watershed-management evaluation monitoring sites, Wisconsin, water years1 1985-93-Continued Start of storm Date (yr/mo/d) Time (24 h) End of storm Date (yr/mo/d) Time (24 h) Precipitation (in.) Streamflow volume (Mft3) Lords Suspendedsolids load (tons) Totalphosphorus load (Ib) Rattlesnake Creek-Continued 90/03/11 90/05/09 90/05/19 90/06/22 90/08/24 90/08/26 91/04/12 91/08/07 91/11/01 91/11/29 92/02/03 92/02/20 92/02/22 92/02/24 92/04/20 92/06/16 92/09/07 92/09/14 92/11/01 92/11/19 92/12/15 93/03/03 93/03/05 93/03/06 93/03/07 93/03/08 93/03/16 93/03/24 93/03/26 93/03/27 93/03/28 93/03/30 93/05/02 93/06/07 93/06/13 93/06/17 93/06/28 93/06/29 90/03/12 90/05/10 90/05/20 90/06/22 90/08/25 90/08/27 91/04/13 91/08/08 91/11/02 91/11/30 92/02/04 92/02/21 92/02/23 92/02/25 92/04/21 92/06/17 92/09/08 92/09/15 92/11/03 92/11/21 92/12/17 93/03/05 93/03/06 93/03/07 93/03/08 93/03/09 93/03/17 93/03/26 93/03/27 93/03/28 93/03/29 93/03/31 93/05/03 93/06/08 93/06/14 93/06/18 93/06/28 93/06/30 s/m

s/m s/m s/m s/m s/m s/m s/m s/m s/m s/m s/m s/m 2,800 1,600 2,100 1,400 4,300 4,400 1,300 1,600 1,000 4,600 7,600 4,400 1,700 1,300 4,000 1,700 1,500 4,700 3,600 4,200 14,000 3,000 4,300 1,700 14,000 2,600 1,100 2,700 4,400

Appendix 1. Storm-load data for rural watershed-management evaluation monitoring sites, Wisconsin, water years1 1985-93~Continued Start of storm Date (yr/mo/d) Time (24 h) End of storm Date (yr/mo/d) Time (24 h) Precipitation (in.) Streamflow volume (Mft3) Loads Suspendedsolids load (tons) Totalphosphorus load db) Rattlesnake Creek-Continued 93/07/05 93/07/08 93/07/10 93/07/17 93/08/14 93/08/15 93/08/18 93/07/06 93/07/09 93/07/12 93/07/18 93/08/15 93/08/16 93/08/19 8,700 8,500 3,000 25,000 24,000 1,700 1,200 'Water year is the 12-month period, October 1 through September 30. The water year is designated by the calendar year in which it ends. Thus, the year starting October 1, 1992 and ending September 30,1993 is called the "1993 water year."

Appendix 2. Storm-load data for urban watershed-management evaluation monitoring sites, Wisconsin, water years 1991-93 [yr, year, mo, month; d, day; h, hour, in., inches; Mfr, million cubic feet; Ib, pounds; s/m, snowmelt; , no data] en Start of storm Date (yr/mo/d) Time (24 h) End of storm Date (yr/mo/d) Time (24 h) Precipitation (in.) Streamflow volume (Mft3) Suspendedsolids (tons) Phosphorus Ob) Loads Total cadmium (Ib) Total copper (Ib) Total lead (Ib) Total zinc (Ib) Lincoln Creek 93/03/22 93/03/24 93/03/31 93/04/02 93/04/19 93/04/29 93/05/22 93/05/30 93/06/04 93/06/07 93/06/08 93/06/14 93/06/17 93/06/19 93/06/30 93/07/03 93/07/05 93/07/08 93/07/13 93/07/25 93/08/09 93/08/15 93/08/30 93/09/13 93/09/20 93/03/24 93/03/26 93/04/02 93/04/05 93/04/21 93/04/29 93/05/24 93/05/31 93/06/05 93/06/08 93/06/08 93/06/14 93/06/18 93/06/20 93/06/30 93/07/04 93/07/06 93/07/09 93/07/14 93/07/25 93/08/09 93/08/16 93/08/31 93/09/15 93/09/21 s/m s/m s/m s/m

Appendix 2. Storm-load data for urban watershed-management evaluation monitoring sites, Wisconsin, water years 1991-93 Continued O3 Start of storm Date (yr/mo/d) Tune (24 h) End of storm Date (yr/mo/d) Time (24 h) Precipitation (in.) Streamflow volume (Mft3) Suspendedsolids (tons) Phosphorus Ob) Loads Total cadmium Ob) Total copper (Ib) Total lead Ob) Total zinc Ob) Menomonee River 91/06/14 91/06/22 91/07/01 91/07/07 91/07/12 91/07/18 91/07/21 91/07/29 91/08/08 91/09/09 91/09/11 91/09/14 91/10/04 91/10/14 91/10/28 92/03/16 92/03/23 92/03/25 92/03/27 92/03/30 92/04/15 92/04/16 92/04/17 92/04/18 92/05/11 91/06/15 91/06/22 91/07/01 91/07/07 91/07/12 91/07/18 91/07/21 91/07/29 91/08/08 91/09/10 91/09/12 91/09/14 91/10/07 91/10/14 91/11/06 92/03/20 92/03/25 92/03/27 92/03/29 92/04/02 92/04/16 92/04/17 92/04/18 92/04/21 92/05/13 s/m s/m s/m s/m s/m s/m s/m s/m s/m

<27

Appendix 2. Storm-load data for urban watershed-management evaluation monitoring sites, Wisconsin, water years1 1991-93 Continued Start of storm Date (yr/mo/d) Time (24 h) End of storm Date (yr/mo/d) Time (24 h) Precipitation (in.) Streamflow volume (Mft3) Suspendedsolids (tons) Loads Total Phosphorus cadmium (Ib) Ob) Total copper (Ib) Total lead Ob) Total zinc (Ib) Menomonee River Continued 92/06/14 92/06/17 92/07/08 92/07/12 92/07/13 92/08/12 92/08/25 92/08/27 92/09/09 92/09/14 92/09/16 92/11/01 92/11/19 92/11/23 92/12/15 93/03/01 93/03/02 93/03/04 93/03/05 93/03/16 93/03/22 93/03/23 93/03/25 93/03/27 93/03/29 92/06/14 92/06/18 92/07/09 92/07/13 92/07/16 92/08/13 92/08/27 92/08/30 92/09/10 92/09/16 92/09/18 92/11/05 92/11/23 92/11/24 92/12/16 93/03/02 93/03/04 93/03/05 93/03/07 93/03/17 93/03/23 93/03/25 93/03/27 93/03/29 93/03/31

s/m s/m s/m s/m s/m s/m s/m s/m s/m s/m <A7 2,300 1,300 79, <19

Appendix 2. Storm-load data for urban watershed-management evaluation monitoring sites, Wisconsin, water years1 1991-93 Continued oo Start of storm Date (yr/mo/d) Time (24 h) End of storm Date (yr/mo/d) Time (24 h) Precipitation (in.) Streamflow volume (Mft3) Suspendedsolids (tons) Phosphorus Loads Total cadmium (lb) Total copper (lb) Total lead Ob) Total zinc (lb) Menomonee River Continued 93/03/31 93/04/14 93/04/19 93/04/29 93/05/22 93/05/30 93/06/04 93/06/19 93/06/30 93/07/05 93/07/08 93/08/30 93/08/30 93/08/31 93/09/13 93/04/01 93/04/19 93/04/23 93/05/01 93/05/24 93/06/01 93/06/13 93/06/21 93/07/01 93/07/06 93/07/13 93/08/30 93/08/30 93/09/03 93/09/14 s/m 2,400 1,200 1,000 1,500 3,300 7,000 5,800 1,500 4,000

1,600 1,000 1,000 Monroe Street detention pond 92/03/16 92/03/22 92/03/28 92/04/08 92/05/11 92/06/17 92/07/02 92/03/16 92/03/22 92/03/29 92/04/09 9TAH/15 92/05/11 92/06/17 92/07/02 s/m

Appendix 2. Storm-load data for urban watershed-management evaluation monitoring sites, Wisconsin, water years1 1991-93-Continued

Start of storm Date (yr/mo/d) Time (24 h) End of storm Date (yr/mo/d) Time (24 h) Loads

Suspended- Total

Streamflow Total Precipitation , solids Phosphorus cadmium Total lead /

volume ,.

copper , . On.) (Mft3) Ob) (ity (lb) Total zinc (lb) Monroe Street detention pond Continued 92/07/08 92/12/30 93/03/03 93/03/05 93/03/06 93/03/07 93/03/08 93/03/22 93/04/14 93/05/01 93/05/22 93/05/30 93/06/02 93/06/07 93/06/24 93/06/30 93/07/05 93/07/09 93/07/25 93/08/15 93/09/13 90/11/21 90/11/27 90/11/27 92/07/08 92/12/30 93/03/03 93/03/05 93/03/07 93/03/07 93/03/08 93/03/23 93/04/16 93/05/02 93/05/24 93/05/30 93/06/03 93/06/07 93/06/25 93/06/30 93/07/05 93/07/09 93/07/25 93/08/15 93/09/14 90/11/21 90/11/27 90/11/27 s/m s/m s/m s/m s/m s/m Nine Springs tributarv storm sewer

Appendix 2. Storm-load data for urban watershed-management evaluation monitoring sites, Wisconsin, water years1 1991-93-Continued Start of storm Date (yr/mo/d) Time (24 h) End of storm Date (yr/mo/d) Time (24 h) Precipitation (in.) Streamflow volume (Mft3) Loads Suspended- Total

solids Phosphorus cadmium Total lead (tons) Ob) Ob) "JJJ* Ob) Total zinc (Ib) Nine Springs tributary storm sewer Continued 90/12/12 91/02/04 91/03/01 91/03/17 91/03/22 91/03/26 91/03/26 91/04/08 91/04/12 91/04/13 91/05/05 91/05/17 91/05/18 91/05/21 91/05/25 91/06/10 91/06/12 91/06/13 91/07/01 91/07/21 91/08/07 91/08/16 91/09/03 91/09/09 90/12/12 91/02/04 91/03/02 91/03/18 91/03/22 91/03/26 91/03/27 91/04/09 91/04/12 91/04/14 91/05/05 91/05/17 91/05/18 91/05/21 91/05/25 91/06/10 91/06/12 91/06/14 91/07/01 91/07/21 91/08/07 91/08/17 91/09/03 91/09/10 s/m s/m

Appendix 2. Storm-load data for urban watershed-management evaluation monitoring sites, Wisconsin, water years1 1991-93-Continued Start of storm Date (yr/mo/d) Time (24 h) End of storm Date (yr/mo/d) Time (24 h) Precipitation (in.) Streamflow volume (Mft3) Suspendedsolids Phosphorus (tons) (Ib) Loads Total cadmium Ob) T°tal Total lead copper (Ib) v ; Total zinc (Ib) Nine Springs tributary storm sewer Continued 91/09/11 91/09/14 91/09/15 91/09/17 91/09/24 91/10/03 91/10/04 91/10/13 91/10/26 91/10/28 91/10/31 91/11/14 91/11/17 91/12/12 92/01/08 92/03/09 92/03/28 92/04/08 92/04/15 92/04/15 92/04/16 92/04/18 92/04/23 92/05/11 91/09/12 91/09/14 91/09/16 91/09/18 91/09/24 91/10/04 91/10/05 91/10/13 91/10/26 91/10/29 91/11/01 91/11/15 91/11/18 91/12/12 92/01/09 92/03/09 92/03/29 92/04/09 92/04/15 92/04/15 92/04/16 92/04/19 92/04/23 92/05/11 <.00079

Appendix 2. Storm-load data for urban watershed-management evaluation monitoring sites, Wisconsin, water years1 1991-93-Continued en to Start of storm Date (yr/mo/d) Time (24 h) End of storm Date (yr/mo/d) Time (24 h) Precipitation (in.) Streamflow volume (Mft3) Suspendedsolids (tons) Loads Total Phosphorus cadmium Ob) (Ib) Total copper (Ib) Total lead Ob) Total zinc (Ib) Nine Springs tributary storm sewer Continued 92/06/17 92/06/24 92/07/02 92/07/08 92/07/13 92/07/16 92/08/29 92/09/06 92/09/09 92/09/16 92/09/17 92/09/18 92/10/15 92/11/01 92/11/12 92/11/19 92/11/22 92/12/15 92/12/29 92/12/30 °3/01/23 93/03/02 93/03/03 93/03/05 92/06/17 92/06/24 92/07/02 92/07/08 92/07/13 92/07/16 92/08/29 92/09/06 92/09/09 92/09/16 92/09/17 92/09/18 92/10/16 92/11/02 92/11/12 92/11/20 92/11/22 92/12/16 92/12/29 92/12/30 03/01/23 93/03/02 93/03/03 93/03/05 s/m s/m s/m s/m .05Q

Appendix 2. Storm-load data for urban watershed-management evaluation monitoring sites, Wisconsin, water years 1991 -93 Continued en co Start of storm Date (yr/mo/d) Tune (24 h) End of storm Date (yr/mo/d) Time (24 h) Precipitation (in.) Streamflow volume (Mft3) Suspendedsolids (tons) Loads Total Phosphorus cadmium (lb) (lb) Total copper (lb) Total lead (lb) Total zinc (lb) Nine Springs tributary storm sewer Continued 93/03/06 93/03/07 93/03/08 93/03/16 93/032 93/03/31 93/04/19 93/05/01 93/05/30 93/06/02 93/06/07 93/064 93/06/30 93/03/06 93/03/07 93/03/08 93/03/16 93/033 93/03/31 93/040 93/05/02 93/05/30 93/06/03 93/06/07 93/06/25 93/06/30 s/m s/m s/m s/m Water year is the 12-month period, October 1 through September 30. The water year is designated by the calendar year in which it ends. Thus, the year starting October 1,1992 and ending September 30, 1993 is called the "1993 water year."

Appendix 3. Quality-assurance/quality-control plan for urban watershed-management evaluation monitoring program, Wisconsin INTRODUCTION The following are quality-assurance/qualitycontrol (QA/QC) procedures that apply to the data-collection projects associated with the urban watershed management evaluation. The purpose of these guidelines is to provide consistent procedures to be used by all field personnel collecting data for the projects. The QA/QC procedures are divided into field procedures and laboratory procedures. All samples are processed in a consistent manner as described in the following sections. All chemical analyses are done by the Wisconsin State Laboratory of Hygiene (WSLH). The WSLH has its own QA/QC procedures (Wisconsin State Laboratory of Hygiene, 1993) which are not discussed herein. The USGS does an inter-laboratory evaluation program semiannually. The WSLH participates in this program, which furnishes a variety of reference samples to accomplish quality-assurance testing of laboratories and to provide an adequate supply of samples that contribute to quality-control programs of participating laboratories. Reports of the results of the standard-reference-sample program and a more detailed description of the program are on file at the USGS office in Madison, Wis. FIELD PROCEDURES Water samples are collected from streams during periods of low flow and high flow and from storm sewers during periods of high flow. Periods of low-flow are at times when the discharge in the stream is not directly affected by precipitation or snowmelt, whereas periods of high flow are at times when the discharge in the stream is increased due to precipitation or snowmelt. Lowflow samples are collected every 2 weeks from April through November and monthly from December through March. Low-flow samples are collected with a DH-81 TMS (trace metal sampler). The outer shell of the sampler is constructed of polypropylene plastic, and the inside of the sampler and the nozzle are made of Teflon. The DH-81 TMS, which can be used with 1/8-, 3/16-, 1/4-, or 5/16-in. nozzles, is suspended from a plastic-coated rod. The 5/16-in. nozzle typically is used to collect samples during low flow. This sampler is used with a 1-L glass mason jar. An equal-width-increment (EWI) sample consists of collecting water at 5 to 10 verticals across the stream. The mason jar, the sairpling apparatus, and the collection jar are rinsed twice with stream water. The mason jar i<? then filled approximately 10 times and emptied into a 10-L glass collection jar. The sample is transported in an iced cooler to the Madison field office for processing. A refrigerated automatic sampler is used to collect samples at high flow. Teflon-lined 3/8-in. sample tubing connects the stream to a peristaltic pump, which pumps water through approximately 3 ft of polyethylene tubing into one of four 10-L glass bottles stored in the refrigerator. The samples are collected on a f ow-composite basis, resulting in an event mean concentration. After each runoff period, the samples are capped and transported in an iced cooler to the Madison field office for processing. SAMPLE PROCESSING Low-flow and high-flow samples are processed identically. The 10-L glass collection bottles are agitated and emptied into a Teflonlined churn splitter. The sample is split into several bottles that have been rinsed with sample water. The sample is then filtered through Gelman capsule filters with a peristaltic pump. The filtered water is collected into two different WSLH sample bottles; sample bottles are rinsed with filtered water before they are fUed. One bottle of sample to be analyzed for organic constituents and the bottle of sample to be analyzed for nutrients are preserved with sulfuric acid. The samples for determination of totf 1 recoverable and dissolved metals are preserved with nitric acid. The sulfuric and nitric acid are acquired from the WSLH. All of these bottles are then transported to the WSLH in an iced cooler. EQUIPMENT CLEANING PROCEDURES The DH-81 TMS sampler, glass mason jar, 10-L glass collection jars, and the churn splitter

Appendix 3. Quality-assurance/quality-control plan for urban watershed-management evaluation monitoring program, Wisconsin-Continued are all cleaned in the same manner. They are initially washed with a nonphosphate soap and tap water. They are then rinsed once with tap water, once with a 10-percent trace-metal-grade hydrochloric acid solution, twice with Milli-Q water (deionized water passed through a carbon filter and a membrane filter), once with methanol, and finally three times with Milli-Q water. Each of the four glass collection bottles in the refrigerated sampler holds 10 to 20 subsamples, depending on the individual site. Before each subsample is collected, the sampler runs through a rinsing procedure. To begin with, the sample tubing is purged. Next, the pump draws streamwater into the point just before the pump head and the sample tubing is purged again. At this point, the rinse cycle is complete and the subsample is collected. USE OF BLANK SAMPLES QA/QC procedures include regular analysis of blank samples to investigate the sampling process for possible sources of contamination. Blank-Sample Processing Sampler blanks are used to evaluate contamination of the entire storm-sampling process, which includes all equipment (automatic sampler, 10-L sample-collection bottles, and churn splitter) and filtering procedures. Milli-Q water is collected into 5-gal glass carboys from the WSLH and is used as reagent water. The blank samples are collected as follows: 1. A set of WSLH sample bottles is first rinsed and then filled directly with Milli-Q water from the carboy. Each bottle in this set is referred to as a "bottle blank before." The samples are preserved with acid as outlined above and refrigerated for possible future use. 2. A 3.8-in.-Teflon-lined extension tube is connected to the end of the sample tubing at the site. The sample tubing is purged. The extension tube is inserted into the carboy, and the Milli-Q water is pumped through the tubing to the point just before it reaches the pump head. The extension tube is taken out of the carboy, and the sample tubing is purged again. rnhe extension tube is inserted back into the carboy, and Milli-Q water is pulled through the sampling system until it fills two 10-L collection bottles. The two bottles are taken out of the sampler. The water f-om these two bottles is used to rinse a set of WSLH sample bottles and then to fill them. Half of the sample is from one of the 10-L bottles and the other half is from the second of the 10-L bottles. Each bottle in this set is referred to as an "ISCO blank." The samples are preserved with ac;d as outlined above and refrigerated for possible future use. 3. Again, a set of WSLH bottles is rinsed and filled with Milli-Q water directly from the carboy. Each bottle in this set is referred to as a "bottle blank after." samples are preserved with acid as outlined above and refrigerated for possible future use. 4. The remaining sample is transported in an iced cooler to the Madison field office. At the office, the sample is emptied into a Teflon-lined churn splitter from which a set of WSLH sample bottles (referred to as "cumulative splitter blanks") rre rinsed and filled for analysis of total an total-recoverable constituents. Some of the water in the churn splitter is filtered and then used to rinse and fill a set of WSLH bottles (referred to as "cumulative file~ blanks") for analysis of filtered constituents. The cumulative splitter blanks and the cumulative filter blanks are preserved with acid as outlined above, chilled, and taen to the WSLH for analysis. If results from the cumulative splitter blanks and cumulative filter blanks are indicative of contamination, then the ISCO blank will be analyzed. If results from the ISCO blanks are indicative of contamination, then the bottle blanks after will be analyzed. If results from the bottle blanks after are indicative of contamination, then the bottle blanks before will be analyzed. If contamination persists with the bottle blanks before, further investigation is needed.

Appendix 3. Quality-assurance/quality-control plan for urban watershed-management evaluation monitoring program, Wisconsin-Continued Evaluation of Blank Samples The field blanks described above are designed to evaluate whether or not samples have been contaminated, and, if so, at what level, through the course of collection, processing, preservation, or transportation. Contamination can be either systematic or erratic. Regardless of the source of contamination or whether the contamination is systematic or erratic, it is useful to define the point where contamination levels are unacceptable. At this point, further investigation is needed to identify and eliminate the sources of contamination. Appendix 4 contains a list of the constituents being monitored, the limit of detection (LOD) and limit of quantification (LOQ) for each constituent, the minimum concentration detected in a water sample during water year 1993 at any of the urban sites, and the concentration where blank sample contamination is suspected and further investigation is warranted. The LOD and LOQ are used for samples analyzed by the WSLH. The LOD is defined as the minimum concentration of a substance that can be measured and reported with 99 percent confidence that the analyte concentration is greater than zero. Near this limit, results are estimated to have an uncertainty in the range of ± 100 percent. The LOQ is defined as the concentration of a substance above which quantitative results may be obtained with a 99 percent degree of confidence. Test results that fall between the LOD and the LOQ are much more uncertain than those that are equal to or greater than the LOQ. A detailed description of the quality assurance program used by the WSLH is on file at the WSLH and the USGS office located in Madison, Wis. The concentration where black sample contamination is suspected is defined in this report as twice the LOD. Sample concentrations near the LOD are very uncertain, so it is estimated that the range where concentrations are at the LOD ± the LOD are too uncertain to support any type of sampling decisions. Above this level, there is an adequate degree of confidence that enough contamination is present to effect stream-sample concentrations. If blank-sample concentrations are consistently found to be greater than twice the LOD, investigation is needed to identify the of contamination, ad after the is identified, whether and how the source of contamination can be eliminated.

Appendix 4. Summary of constituent concentrations indicative of contamination in sample blanks for urban watershed-management evaluation monitoring sites, Wisconsin [mg/L, milligrams per liter; ug/L, micrograms per liter; --, undefined] Constituent BOD5, total COD, low level Calcium, dissolved Magnesium, dissolved Hardness as CaCOa, dissolved Alkalinity Chloride, dissolved, high range Solids, suspended Solids, total Nitrate + nitrite Nitrogen ammonia Phosphorus, total Phosphorus, dissolved, low range Cadmium, dissolved Cadmium, total recoverable Copper, dissolved Copper, total recoverable Zinc, dissolved Zinc, total recoverable Lead, dissolved Lead, total recoverable Limit of detection 0.3 mg/L 5.0 mg/L .02 mg/L .02 mg/L 6.0 mg/L 1.0 mg/L .lmg/L 2.0 mg/L 10. mg/L .007 mg/L .005 mg/L .008 mg/L 2.0 ug/L .04 ug/L .16 ug/L 1.0 ug/L 1.0 ug/L 10. ug/L 10. ug/L 1.0 ug/L 1.2 ug/L Limit of Lowest water sample quantification concentration, water year 1993 16. mg/L .05 mg/L .07 mg/L

A mg/L .03 mg/L .019 mg/L .031 mg/L 5.0 ug/L .15 ug/L .52 ug/L 3.0 ug/L 3.0 ug/L 40. ug/L 40. ug/L 3.0 ug/L 4.0 ug/L <1.0 mg/L <5.0 mg/L 13. mg/L 2. mg/L 43. mg/L 51 mg/L 5.0 mg/L 2. mg/L 160. mg/L .22 mg/L .017 mg/L .03 mg/L 2.0 ug/L <.04ug/L <.2ug/L 1.8 ug/L 4.0 ug/L <10. ug/L <10. ug/L <1.0ug/L <3.0 ug/L Concentration where contamination is suspected 0.6 mg/L 10. mg/L .04 mg/L .04 mg/L 12. mg/L 2. mg/L .2 mg/L 4.0 mg/L 20. n-e/L .014 mg/L .010 mg/L .016 mg/L 4.0 ug/L .Of ug/L .3? ug/L 2.0 ug/L 2.0 ug/L 20. |i-?/L 20. U3/L 2.0 ug/L 2.4 ug/L