Scientific Investigations Report 2004–5156

In cooperation with the Dane County Land Conservation Department and the Wisconsin Department of Natural Resources

# Hydrologic, Ecologic, and Geomorphic Responses of Brewery Creek to Construction of a Residential Subdivision, Dane County, Wisconsin, 1999–2002

## Tables

1 Dane County Land Conservation Department, Madison, Wisconsin

2 Wisconsin Department of Natural Resources, Madison, Wisconsin

## Conversion Factors and Abbreviated Water-Quality Units

Multiply By To obtain
Length
inch (in) 2.54 centimeter (cm)
foot (ft) 0.3048 meter (m)
mile (mi) 1.609 kilometer (km)
Area
acre 4,047 square meter (m2)
acre .4047 hectare (ha)
acre .004047 square kilometer (km2)
square ft (ft2) .09290 square meter (m2)
square mile (mi2) 259.0 hectare (ha)
square mile (mi2) 2.590 square kilometer (km2)
Volume
gallon (gal) 3.785 liter (L)
cubic foot (ft3) .02832 cubic meter (m3)
acre-foot (acre-ft) 1,233 cubic meter (m3)
Flow rate
cubic foot per second (ft3/s) .02832 cubic meter per second (m3/s)
gallon per minute (gal/min) .06309 liter per second (L/s)
inch per hour (in/h) .0254 meter per hour (m/h)
Mass
pound, avoirdupois (lb) .4536 kilogram (kg)
ton, short (2,000 lb) .9072 megagram (Mg)
foot squared per day (ft2/d) .09290 meter squared per day (m2/d)

Temperature in degrees Celsius (°C) may be converted to degrees Fahrenheit (°F) as follows: °F=(1.8x°C)+32

Temperature in degrees Fahrenheit (°F) may be converted to degrees Celsius (°C) as follows: °C=(°F-32)/1.8

Concentrations of chemical constituents in water are given either in milligrams per liter (mg/L) or micrograms per liter (µg/L).

## Abstract

The U.S. Geological Survey (USGS), in cooperation with the Dane County Land Conservation Department (LCD) and the Wisconsin Department of Natural Resources (DNR), investigated the instream effects from construction of a residential subdivision on Brewery Creek in Dane County, Wisconsin. The purpose of the investigation was to determine whether a variety of storm-runoff and erosion-control best-management practices (BMPs) would effectively control the overall sediment load, as well as minimize any hydrologic, ecologic, and geomorphic stresses to Brewery Creek.

Stormwater volumes decreased 60 percent from the preconstruction phase to the land-disturbance phase and slightly increased (9 percent) from the land-disturbance phase to the home-construction phase. The stormwater volumes were applied to total solids and total suspended solids concentrations to compute a solids load for each contaminant. Total and suspended solids load indicated a similar trend from preconstruction to land-disturbance phases with decreases of 52 and 72 percent, respectively. Both total and suspended solids load continued to decrease in the transition from land-disturbance to home-construction phases, by 22 and 37 percent, respectively. However, because of variability in the data, statistically there was no change in the magnitude of difference between the upstream and downstream solids load from one phase of construction to the next at the 90-percent confidence level.

Other physical, biological, and ecological surveys including macroinvertebrates, fish, habitat, and geomorphology were done on segments of Brewery Creek affected by the study area. Macroinvertebrate sampling results (Hilsenhoff Biotic Index value, or HBI), on Brewery Creek ranged from "very good" to "good" water-quality with no appreciable differences during any phase of construction activity. Results for fish-community composition, however, were within the "poor" range (Index of Biotic Integrity value, or IBI) during each year of testing. A general absence of intolerant species, with the exception of brown trout, reflects the low IBI values. Habitat values did not change significantly from preconstruction to postconstruction phases. Although installation of a double-celled culvert in Brewery Creek most likely altered the width-to-depth ratio in that reach, the overall habitat rating remained "fair". Fluvial geomorphology classifications including channel cross sections, bed- and bank-erosion surveys, and pebble counts did not indicate that stream geomorphic characteristics were altered by home-construction activity in the study area. Increases in fine-grained sediment at various cross sections were attributed to instream erosion processes, such as bank slumping, rather than increases in sediment delivery from the nearby construction site.

## Introduction

Controlling nonpoint sources of water contamination has been a major focus of the regulatory community in recent years. Because of the past and current successes in controlling contamination from point sources, contamination from nonpoint sources (including sediment deposition, erosion, contaminated runoff, hydrologic modifications that degrade water quality, and other diffuse sources of contaminants) is now the largest cause of water-quality impairment in the United States (U.S. Environmental Protection Agency, 2001).

Conversion of rural and agricultural lands to developed urban areas is a leading contributor of nonpoint-source pollution. Urban development generates numerous contaminants that are associated with the activities of dense populations. Urban development also increases the amount of impervious surface in a watershed as farmland, forests, and meadowlands with generally high infiltration characteristics are converted into buildings with rooftops, driveways, sidewalks, roads, and parking lots with virtually no capacity to absorb stormwater. When stormwater and snowmelt runoff wash over these impervious areas, the runoff picks up contaminants along the way while gaining speed and volume, because it does not have the capacity to disperse and filter into the ground. The results are stormwater flows that are higher in volume, contaminant load, and temperature than the flows in less developed areas, which generally have more natural vegetation and soil to filter the runoff (U.S. Environmental Protection Agency, 1997).

Although water quality across the country has improved since passage of the Clean Water Act in 1972, various challenges still remain. In 2000, water-quality assessments conducted by States indicated that 39 percent of assessed stream miles, 45 percent of assessed acres of lakes, and 51 percent of assessed estuary areas failed to meet criteria for one or more designated uses. The top causes of impairment in assessed stream miles were siltation, nutrients, bacteria, metals (primarily mercury), and oxygen-depleting substances. Pollution from urban and agricultural land that is transported by precipitation and runoff was found to be the leading source of impairment (U.S. Environmental Protection Agency, 2002).

The U.S. Geological Survey (USGS), in cooperation with the Dane County Land Conservation Department (LCD) and the Wisconsin Department of Natural Resources (WDNR), investigated the instream water-quality effects from construction of a residential subdivision on Brewery Creek, Dane County, Wis. The purpose of the investigation was to determine whether storm-runoff and erosion-control best-management practices (BMPs) would effectively control the overall sediment load, as well as minimize any physical, biological, and ecological stresses to Brewery Creek.

Few previous studies have assessed the capacity of erosion and sediment controls and stormwater-management practices to prevent degradation of receiving waters in urbanizing areas. Even fewer studies have been multiparameter investigations, integrating water-quality observations with evaluations of stream physical habitat and biological quality. This investigation paired water-quality analyses with physical habitat, stream geomorphology, and biological indices to evaluate the capacity of selected management techniques to prevent degradation of a receiving stream.

This study has relevance at both national and local levels. At the national level, the investigation could provide necessary background data on water-quality impairments related to construction-site runoff. This, in turn, would help facilitate the implementation of Phase II of the USEPA National Pollution Discharge Elimination System (NPDES) standard for pollution control on construction sites of less than 5 acres. At the local level, county officials are mandating construction-site erosion-control standards throughout Dane County. The objective of this investigation was to provide evidence of the effects that construction has on hydrology, ecology, and morphology of receiving waters.

### Purpose and Scope

This report describes the methods used in and the results from the Brewery Creek study. An upstream-downstream (above-and-below) experimental design was used to isolate the pollutant loads coming from the construction site. Automated, intensive stream-water sampling took place during storm-runoff periods in three different phases on the project: preconstruction (October 1999 to April 2001), land disturbance (May 2001 to March 2002), and home construction (April 2002 to September 2002). Concentrations of total solids and total suspended solids in stream-water samples were used to compute storm loads for each contaminant contributed to Brewery Creek during each phase. In addition to water quality and quantity, other physical and biological data were analyzed to determine the effectiveness of storm-runoff and erosion controls in protecting the integrity of Brewery Creek. Geomorphology classifications, including bed- and bank-material characterizations, were done at intervals throughout the study period. Stream temperatures were recorded at 15-minute intervals during each phase of the project. Annual fish surveys were done to determine species composition and density. Finally, macroinvertebrate and habitat data were collected at various intervals to assess the overall health of the stream.

### Description of Study Area

Brewery Creek is in the Black Earth Creek watershed, in northwestern Dane County (fig. 1). The drainage area of 10.5 mi2 at the downstream end of the study area includes 2.8 mi2 of noncontributing area. The stream is 6.1 mi long from the downstream station to the stream headwaters; 0.5 mi is within the study area. The stream has been channelized in places, with the upper reaches of the stream last being dredged in 1976. The stream bed material is mostly soft silt and clay. Brewery Creek flows through outwash and alluvium composed of sandstone with some shale; most of the bedrock in the watershed is dolomite (Graczyk and others, 2003). The soils of Brewery Creek Watershed are predominantly silt loams that are poorly drained in valley bottoms and highly erodible in the uplands (Glocker and Patzer, 1978). Brewery Creek is a warm-water stream that maintains a forage fish population (Wisconsin Department of Natural Resources, 1989). Although classified as a warm-water stream, the stream does support cold-water species and is a candidate to be reclassified as a cold-water stream by the WDNR (Wisconsin Department of Natural Resources, 1989). The largest land-use categories in the Brewery Creek Watershed are agriculture, at 57 percent, and woodland, at 22 percent (Graczyk and others, 2003).

The St. Francis residential subdivision includes single-family lot development over approximately 72 acres. It lies in the southernmost part of the Brewery Creek Watershed and represents approximately 4 percent of the basin area (fig. 2). The area had previously been used for corn and soybean production. An established vegetated stream buffer varies from 30 to 100 ft in width on either side of the stream.

### Erosion Control and Stormwater Management at the Subdivision Site

Since August 22, 2002, all municipalities in Dane County, Wisconsin have been required to meet the requirements of the Dane County Erosion Control and Stormwater Management Ordinance. In Dane County, all developments disturbing more than 4,000 ft2 are required to implement an erosion-control plan. Stormwater-management plans are required when 20,000 ft2 or more of impervious surface is created. (Chapter 14, Dane County Code of Ordinances). During the general permitting of the St. Francis subdivision, these regulations were not applicable because the permit was issued in April 2001. However, the developer agreed to design and implement an erosion-control and stormwater-management plan that would meet or exceed requirements in the proposed Dane County Ordinance.

The St. Francis subdivision employed a variety of BMPs constructed for erosion control and stormwater management. The erosion-control practices were designed to meet the maximum allowable cumulative soil loss of 7.5 ton/acre/yr. Practices included installing silt fence reinforced with straw bales, maintaining vegetative buffers (fig. 3), sequencing construction, deep tilling to minimize compaction, temporary seeding of soil stockpiles, protecting inlets, emplacing stone tracking pads, and building temporary earthen berms.

Stormwater-management practices were designed and implemented in accordance with the water-quality and -quantity standards under development in the Dane County Erosion Control and Stormwater Management Ordinance. The applicable standards included the following:

• Maintaining the predevelopment peak-runoff rates for the 2-year and 10-year, 24-hour storms, and safely passing the 100-year flood.
• Discharging to a stable outlet carrying the designed flows at a nonerosive velocity.
• Retaining all soil particles greater than 5 microns.
• Directing runoff from downspouts, driveways, and other impervious areas to pervious areas.
• Including provisions and practices to reduce the temperature of runoff to the receiving waters.

Stormwater-management practices included grassed swales and boulevards for infiltration and storage of runoff, reduced street widths to minimize impervious cover, protection of present woodlands, two detention and infiltration basins with stone cribs for thermal protection, maintenance of stream buffers, and use of available parkland and open space for runoff storage and infiltration. Figure 4 highlights some of the BMPs used in the development of the study area.

The site was designed to maximize infiltration of runoff on the basis of predevelopment soil and permeability rates. Surface runoff is diverted from impervious surfaces to one or more BMPs for temporary storage and infiltration. Runoff first enters grassed swales in the street medians (fig. 4). The swales were designed to infiltrate stormwater over a period of 24 hours. During periods of intense runoff, excess water in the swale enters a conveyance system that directs runoff to a larger infiltration basin, where it is temporarily stored and allowed to infiltrate. Each system is designed to reduce water quantity and improve water quality before runoff enters Brewery Creek. Vegetated buffers were left intact during site grading and plot construction to provide additional water-quality benefit.

## Methods of Data Collection

Data collection in the study area involved water-quantity and sediment measurement at two monitoring stations and habitat, biologic, and geomorphic data near each station. Locations of the data-collection stations are shown in figure 5.

### Water-Quantity, Precipitation, and Water-Quality Measurement

A stream-monitoring station had been established at the downstream site in 1984 and was active from 1984 to 1986; 1989 to 1998, and May 1999 to September 2002. The upstream monitoring station was established in October 1999. The location of each station in relation to the study area is shown in figure 5. Although the upstream station appears to be near the center of the study area, all construction activity was confined to the area between the upstream and downstream stations for the duration of the study. Future development has been planned beyond the upstream station. Each station continuously measured stream levels and water temperature, and event-based water samples also were collected. Water-level measurements were recorded in 15-minute increments during periods of base flow and 5-minute increments during storm events.

A storm event was defined as a period of precipitation bracketed by 6 hours or more of no precipitation. In some cases, storm events were defined as a period of precipitation bracketed by 12 hours or more of no precipitation; these events typically were the result of stormwater runoff continuing beyond a 6-hour period of no precipitation, followed by a second burst of rainfall causing additional runoff.

Water Quantity. Changes in stream levels were measured with a bubble-gage system and pressure transducer. Stream levels were then converted to a discharge rate by use of a field-verified rating table. A V-notch weir was added to the upstream station to gain added sensitivity of measured discharge for small fluctuations in water level, whereas the downstream channel cross section provided sufficient accuracy.

Precipitation. Precipitation was measured at the downstream station with a tipping-bucket raingage. Because of the close proximity to the upstream station, all precipitation data were collected at a single station. Precipitation depths, intensities, and erosivity indices were computed for all storm events except snowmelt. See tables A1 through A3 in the appendix for precipitation data. Intensities are reported in 5-, 10-, 15-, 30-, and 60-minute increments.

Water Quality. Stream temperature was measured with a Teflon-shielded thermocouple at a single point in the water column at each stream-monitoring station.

Automated water samplers at each station collected samples for water-quality analyses. Sample collection was activated by a rise in stream level during a storm event. Once a stream-level threshold was exceeded, typically a rise of 0.10 ft above base-flow level, the volume of water passing the station was measured and accumulated at 1-minute increments until a volumetric threshold was reached. At that point, the sampler collected a discrete water sample and the volumetric counter was reset. The process was repeated until the stream level receded below the threshold.

These flow-weighted samples were collected and composited into a single water sample, then split and processed for analysis. A Teflon-coated, stainless-steel churn splitter was used to composite and split samples. Processed samples were placed on ice and taken to the Wisconsin State Laboratory of Hygiene (WSLH) within 48 hours after runoff cessation for determination of concentrations of total and suspended solids. Because each discrete sample was composited into a single event sample, the resulting concentration represents the event mean concentration (EMC). In some cases, individual discrete samples were submitted to the laboratory to gain a better understanding of concentration variations during a storm event. Figure 6 illustrates how discrete samples were acquired over a single storm event.

Solids loads were computed by multiplying the EMC by the total volume of the storm event and a constant for unit conversion. For those events in which discrete-sample concentrations were used rather than an EMC, continuous streamflow and instantaneous concentration data were used to estimate loads of total and suspended solids. In this case, loads were computed by summing the product of streamwater-sample concentration and streamflow rate for that storm-runoff period (Porterfield, 1972).

To ensure sample integrity, field and sample-processing equipment blanks were collected at the upstream and downstream stations. Blank samples were obtained by drawing deionized water through the suction line and sampler into a collection bottle. The Teflon sample line and automatic sampler were not cleaned before obtaining blank samples. Blank water collected in the sample bottle was then run through the Teflon-lined churn splitter into laboratory-prepared sample bottles. Samples were placed on ice and delivered to the WSLH for analysis. Deionized blank water was also used to isolate individual elements of the sampling process from source to delivery. These samples were not delivered to the WSLH unless erroneous concentrations were found in the original blank sample. Blank-sample results are detailed in table A4 in the appendix. A significant concentration of total and suspended solids was detected in the upstream blank sample collected in August 1999. This blank sample may have been compromised by stream water entering the sample tubing while the blank sample was being acquired. An additional blank sample was acquired as an added quality measure. The results of that sample fell within acceptable limits.

Sample-collection bottles were cleaned with a nonphosphate detergent, tapwater rinse, and hydrochloric acid rinse and then were air-dried. Clean bottles replaced soiled bottles upon collection of the samples and remained in the sampler housing until the next runoff event. A Teflon-lined churn splitter was rinsed with deionized water before sample processing.

Replicate samples were submitted to verify reproducibility with automatically collected samples. Replicate samples were checked for precision on the basis of a relative percent difference (RPD). Manual samples were collected periodically to verify reproducibility with corresponding stream equal-width-increment (EWI) samples (Ward and Harr, 1990). Manual samples were also checked for precision on the basis of a relative percent difference. Total and suspended solids RPD values fell below the 20-percent-precision criteria for both upstream and downstream replicate samples. However, RPD values exceeded the 20-percent criteria on suspended solids at the upstream station and downstream station in August 2002. Replicate and manual sample results are listed in table A4 in the appendix.

### Macroinvertebrate- and Fish-Community Assessment

Macroinvertebrate communities were sampled in the spring and (or) fall beginning in October 1999. A D-frame net was used to sample riffle habitats in Brewery Creek at Brewery Road (fig. 5). An additional site located on Brewery Creek at Highway 14 (fig. 7) was also sampled because of its extensive historical macroinvertebrate data record for comparison. Samples were submitted to the Biomonitoring Laboratory at the University of Wisconsin at Stevens Point for processing and identification. The semi-quantitative methodology used in the study for sampling and biotic index calculation was the Hilsenhoff Biotic Index, or HBI (Hilsenhoff, 1987), which is based on sensitivity of various aquatic insects and crustaceans to organic contamination. The HBI water-quality scale ranges from 0 to 10, with 0 indicating best possible water quality and 10 the worst.

Fish communities were sampled at least once per year beginning in October 1999. A towed barge with two electrodes was used to sample 160 m of stream above Brewery Road (fig. 5). The sampling methodology and Coldwater Index of Biotic Integrity, or IBI, calculation used for assessing the environmental health of trout streams was developed by Lyons and others (1996). The Coldwater IBI is based on variable tolerances of different fish species to environmental degradation. Scores range from 0 (worst) to 100 (best). The presence of numerous trout, intolerant species, and numerous species adapted to cold temperatures score the highest and indicate favorable stream conditions.

### Habitat Assessment

Physical characteristics of the stream at both sites were measured to document present conditions. Measurements included stream width, stream depth, depth of fines, bank erosion, substrate type and amount, and cover for fish; all were recorded at 48 transects at each "habitat station" (H1 and H2), a stream reach whose length is 36 times the mean stream width (fig. 5).

Surveys also were done during summer 2003 to determine whether construction had an effect on fish habitat. The surveys followed methods outlined in "Guidelines For Evaluating Fish Habitat in Wisconsin Streams," (Simonson and others, 1994). Qualitative ratings have been established to characterize the physical habitat available for fish. The habitat scores range from 0 to 100, with 0 indicating the worst habitat for fish and 100 being optimal.

### Geomorphic Assessment

The segment of Brewery Creek investigated for this study included both straightened channel and natural channel (pools, riffles, and runs). The upper reaches of the creek have been hydraulically manipulated by past dredging for agricultural purposes.

Various methods were used to determine physical stream characteristics and any subsequent changes resulting from construction activity. Stream classifications, channel cross sections, pebble counts, and bed- and bank-erosion surveys were done according to methods outlined in "Stream Channel Reference Sites: An Illustrated Guide to Field Technique" (Harrelson and others, 1994). Stream classifications were done in 2001 and 2003, whereas channel cross sections and bed- and bank-erosion surveys were completed during 2001 to 2003.

Fourteen cross sections were established with permanent monuments placed in representative locations (fig. 5). Monuments consisted of 4-ft sections of rebar anchored into the ground with 4-in diameter pvc tubes filled with concrete (fig. 8). Annual surveys were conducted by use of a surveyor's level. Survey points included monuments, top of bank, bankfull, bank pins, and water levels.

Bank Pins. Bank pins consisting of 4-ft sections of 3/8-in. rebar were inserted horizontally into the streambank at most permanent cross sections at or slightly above bankfull in fall 2001. Bank pins were set flush with the bank and were intended to measure subtle changes in erosion and deposition to the banks. Measurements of the amount of material either deposited or the amounts of the bank pin exposed (erosion), in millimeters were made in September 2002 and October 2003 and noted differences as either deposition or erosion. If possible, pins were reset to "0" or flush with the bank. Because of severe slumping, this was not possible in most locations. Data were recorded and used for comparison at all cross sections where bank pins were installed.

Stream Classification. Two segments of Brewery Creek were classified using Rosgen's Stream Classification System (Rosgen, 1996). Cross-sections 1 (the furthest upstream site in the channelized reach) and 9 (meandering reach) were classified in October 2001 and repeated in October 2003 to compare different reaches of stream. Stream variables used in the classification procedure include slope, sinuosity, width/depth ratio, entrenchment ratio, and dominant bed material.

Wolman Pebble Count. The Wolman Pebble Count Procedure was used to characterize the composition of the streambed at seven locations. Pebble counts were done in 2001, 2002, and 2003. Selected reaches were sampled (step-toe procedure) from bankfull to bankfull in a random fashion. A minimum of 100 samples were recorded per location. Particles were tallied according to the Wentworth size classes. Particles larger than sand (greater than 2 mm) were measured along the intermediate axis (fig. 9) and recorded under the appropriate size class. Data were plotted annually by size class and frequency.

## Hydrologic Response

Data on runoff volume, solids load, and EMC for each storm event are listed in tables 1 and 2; a statistical summary of runoff volume and solids results for each phase of the study is given in table 3. Temperature data are listed in table 4.

[** - Discrete samples only; mg/L, milligrams per liter]

Brewery Creek upstream
Sampled runoff event Storm information Average solids loads Event mean concentration (EMC)
Start date Construction phase Precip. depth (inches) Runoff volume (cubic feet) Total solids (tons) Suspended solids (tons) Total solids (mg/L) Suspended solids (mg/L)
02/22/00 Preconstruction snowmelt 3,574,500 72.5 36.9 650** 331**
04/19/00 Preconstruction 2.13 1,531,300 27.5 6.6 576** 138**
05/17/00 Preconstruction 5.04 6,116,800 307.4 263.5 1,610 1,380
05/30/00 Preconstruction 5.07 17,508,000 1,032.9 890.8 1,890 1,630
06/04/00 Preconstruction .88 807,500 13.8 3.9 548 156
06/13/00 Preconstruction 2.70 4,491,300 171.0 136.0 1,220 970
07/02/00 Preconstruction .99 661,900 19.8 13.7 960 664
07/10/00 Preconstruction .73 322,200 4.4 .8 438 75
08/05/00 Preconstruction 1.81 422,100 5.8 .7 438 56
08/17/00 Preconstruction .96 268,900 4.2 .2 496 17.5
09/11/00 Preconstruction 1.08 625,000 9.4 1.1 484 54
09/22/00 Preconstruction .64 507,900 7.9 .6 496 35
04/11/01 Preconstruction .73 392,100 7.5 2.9 546 136
05/10/01 Land-disturbance .75 152,900 2.5 .7 596 157
05/21/01 Land-disturbance 2.02 1,334,800 29.2 14.8 662 298
05/23/01 Land-disturbance .54 368,100 6.2 .9 518 68
06/05/01 Land-disturbance .73 268,700 4.6 1.4 534 101
06/11/01 Land-disturbance 2.07 1,937,800 57.1 38.7 944 640
08/01/01 Land-disturbance 9.56 22,855,400 694.4 570.2 974** 800**
08/22/01 Land-disturbance .20 119,300 1.9 .1 502 36
08/25/01 Land-disturbance 1.31 365,400 5.5 .9 494 57
09/17/01 Land-disturbance .57 321,800 5.2 .5 522 48
09/19/01 Land-disturbance 1.06 662,900 10.0 3.1 498 114
09/23/01 Land-disturbance 1.73 1,424,500 27.0 15.5 584 262
10/22/01 Land-disturbance 1.17 462,700 7.6 1.7 520 90
11/24/01 Land-disturbance .82 342,000 6.7 1.9 574 119
12/12/01 Land-disturbance .58 291,900 5.9 1.9 586 141
02/18/02 Land-disturbance 1.68 1,878,600 37.0 14.3 600 199
03/08/02 Land-disturbance .45 1,066,200 20.2 7.5 592 206
04/07/02 Home-construction 1.35 1,057,500 18.4 5.1 558 154
04/18/02 Home-construction .66 220,400 4.1 1.1 602 153
04/24/02 Home-construction .39 177,300 2.9 .6 530 103
04/27/02 Home-construction .70 556,900 9.3 1.3 534 74
05/01/02 Home-construction .56 271,600 4.5 .8 526 92
05/09/02 Home-construction .78 461,800 11.5 6.4 796 442
05/11/02 Home-construction .83 637,600 14.6 7.8 696 337
05/28/02 Home-construction .30 134,000 2.2 .3 522 75
06/03/02 Home-construction .24 517,400 23.1 18.7 1,430 1,160
06/04/02 Home-construction .36 1,206,600 22.6 8.6 600 228
07/22/02 Home-construction 1.16 251,000 3.90 .4 498 47
08/11/02 Home-construction 1.74 431,300 6.1 1.5 460 94
08/21/02 Home-construction .93 479,700 7.7 .6 516 38
09/02/02 Home-construction 1.05 312,400 5.3 .7 540 67

[** - Discrete samples only; mg/L, milligrams per liter]

Brewery Creek downstream
Sampled runoff event Storm information Average solids loads Event mean concentration (EMC)
Start date Construction phase Precip. depth (inches) Runoff volume (cubic feet) Total solids (tons) Suspended solids (tons) Total solids (mg/L) Suspended solids (mg/L)
02/22/00 Preconstruction snowmelt 3,862,800 87.0 50.5 722** 419**
04/19/00 Preconstruction 2.13 1,755,700 32.1 8.6 586** 157**
05/17/00 Preconstruction 5.04 7,077,700 291.6 229.8 1,320 1,040
05/30/00 Preconstruction 5.07 18,098,100 1,050.8 909.5 1,860 1,610
06/04/00 Preconstruction .88 908,000 13.0 1.4 458 48
06/13/00 Preconstruction 2.70 4,838,800 179.7 149.5 1,190 990
07/02/00 Preconstruction .99 662,600 20.7 15.2 1,000 736
07/10/00 Preconstruction .73 343,200 5.1 .7 476 62
08/05/00 Preconstruction 1.81 521,300 8.0 2.1 490 130
08/17/00 Preconstruction .96 308,800 4.7 .2 486 24
09/11/00 Preconstruction 1.08 766,900 11.9 2.0 498 82
09/22/00 Preconstruction .64 580,600 9.2 .8 506 42
04/11/01 Preconstruction .73 451,000 9.4 4.0 614 218
05/10/01 Land-disturbance .75 150,400 1.9 .6 624 205
05/21/01 Land-disturbance 2.02 1,383,500 37.3 22.5 864 522
05/23/01 Land-disturbance .54 415,800 6.6 1.1 512 84
06/05/01 Land-disturbance .73 329,100 6.1 1.8 566 124
06/11/01 Land-disturbance 2.07 2,104,700 96.6 80.8 1,470 1,230
08/01/01 Land-disturbance 9.56 24,816,900 930.2 802.0 1,202** 1,036**
08/22/01 Land-disturbance .20 147,300 1.8 .2 488 49
08/25/01 Land-disturbance 1.31 352,700 5.8 1.4 518 94
09/17/01 Land-disturbance .57 342,400 5.7 .5 534 49
09/19/01 Land-disturbance 1.06 696,400 11.4 2.8 522 117
09/23/01 Land-disturbance 1.73 1,535,200 33.7 20.3 654 325
10/22/01 Land-disturbance 1.17 558,200 9.5 2.8 526 110
11/24/01 Land-disturbance .82 341,600 6.4 1.7 576 136
12/12/01 Land-disturbance .58 273,300 4.9 1.3 536 102
02/18/02 Land-disturbance 1.68 1,943,400 40.8 17.5 632 234
03/08/02 Land-disturbance .45 1,070,500 19.9 6.8 582 188
04/07/02 Home-construction 1.35 1,096,300 18.3 4.2 536 123
04/18/02 Home-construction .66 240,300 4.2 1.1 562 144
04/24/02 Home-construction .39 189,200 3.0 .5 506 79
04/27/02 Home-construction .70 568,300 8.8 1.0 496 57
05/01/02 Home-construction .56 292,800 4.4 .6 484 70
05/09/02 Home-construction .78 511,800 11.7 6.5 734 409
05/11/02 Home-construction .83 727,500 16.7 8.9 688 332
05/28/02 Home-construction .30 152,700 2.4 .4 512 76
06/03/02 Home-construction .24 571,800 24.5 20.2 1,370 1,130
06/04/02 Home-construction .36 1,398,600 25.9 11.0 594 251
07/22/02 Home-construction 1.16 281,800 4.4 .8 498 96
08/11/02 Home-construction 1.74 525,500 7.3 2.6 452 127
08/21/02 Home-construction .93 558,000 8.6 .7 494 42
09/02/02 Home-construction 1.05 380,000 5.9 .7 490 50

Construction phase
Statistic Pre-construction Land-disturbance Home-construction
DOWNSTREAM SITE
Volume (cubic feet x 106)
Mean 3.09 2.28 .54
Median .77 .49 .52
Maximum 18.10 24.82 1.40
Minimum .31 .15 .15
Coeff. of variation 1.61 2.65 .66
Total solids (tons)
Mean 132.5 76.2 10.4
Median 13 8 8
Maximum 1050.75 930.2 25.93
Minimum 4.69 1.81 2.44
Coeff. of variation 2.18 3.01 .75
Suspended solids (tons)
Mean 105.71 60.25 4.22
Median 3.95 2.26 1.05
Maximum 909.52 802 20.17
Minimum .23 .18 .36
Coeff. of variation 2.38 3.3 1.36
UPSTREAM SITE
Volume (cubic feet x 106)
Mean 2.86 2.12 .48
Median .66 .42 .45
Maximum 17.51 22.86 1.21
Minimum .27 .12 .13
Coeff. of variation 1.67 2.63 .66
Total solids (tons)
Mean 129.54 57.57 9.73
Median 13.81 7.16 6.91
Maximum 1032.88 694.4 23.09
Minimum 4.16 1.87 2.18
Coeff. of variation 2.21 2.96 .74
Suspended solids (tons)
Mean 104.43 42.12 3.83
Median 3.93 1.9 1.17
Maximum 890.79 570.2 18.73
Minimum .15 .13 .31
Coeff. of variation 2.38 3.35 1.36

[°C, degrees Celsius; Std Dev, standard deviation]

Temperature °C
Preconstruction Land disturbance Home construction
Event date Upstream Downstream Event date Upstream Downstream Event date Upstream Downstream
02/22/00 4.1 4.3 05/10/01 14.4 15.1 04/07/02 6.3 6.4
04/19/00 9.0 9.5 05/21/01 13.4 13.7 04/18/02 12.0 13.1
05/17/00 11.4 11.7 05/23/01 11.3 11.5 04/24/02 9.2 9.9
05/30/00 16.0 16.2 06/05/01 10.8 10.8 04/27/02 7.0 7.1
06/04/00 13.0 13.3 06/11/01 16.8 17.5 05/01/02 9.1 9.4
06/13/00 15.8 16.1 08/01/01 21.0 21.9 05/09/02 11.8 12.5
07/02/00 17.1 17.7 08/22/01 15.2 16.4 05/11/02 8.6 8.9
07/10/00 18.0 18.8 09/17/01 13.3 13.7 05/28/02 15.5 15.7
08/05/00 16.1 16.6 09/19/01 14.1 14.3 06/03/02 11.8 12.0
08/17/00 14.6 15.1 09/23/01 13.1 13.3 06/04/02 13.0 13.2
09/11/00 16.1 16.5 10/22/01 9.8 10.0 07/22/02 16.8 18.0
09/22/00 11.8 12.0 11/24/01 9.8 9.9 08/11/02 16.8 18.0
04/11/01 8.8 9.1 12/12/01 6.0 6.0 08/21/02 15.6 16.4
02/18/02 3.8 4.0 09/02/02 15.7 16.4
03/08/02 3.5 3.7
Mean 13.2 13.6 Mean 11.7 12.1 Mean 12.1 12.6
Median 14.6 15.1 Median 13.1 13.3 Median 11.9 12.8
Std Dev 4.0 4.1 Std Dev 4.8 5.0 Std Dev 3.6 3.9
Maximum 18.0 18.8 Maximum 21.0 21.9 Maximum 16.8 18.0
Minimum 4.1 4.3 Minimum 3.5 3.7 Minimum 6.3 6.4

Large differences in rainfall patterns between each phase could potentially bias the results of data analyses. To determine whether each construction phase differed with respect to rainfall depth and intensity (in the form of an erosivity index), the Kruskal-Wallis test was used. Precipitation depth and intensities indicated a nonnormal distribution. The Kruskal-Wallis test checks for a difference between the medians of independent samples for nonnormal datasets. No significant differences between the preconstruction and land-disturbance phases were detected at the 90-percent confidence level. Similarly, no significant differences were detected between the land-disturbance and home-construction phases. Therefore, any differences between the downstream and upstream stations between study phases are not likely because of differences in rainfall patterns.

On the whole, the downstream-upstream experimental design worked well at documenting the effects of the BMP systems from preconstruction through the home-construction phases. This design could continue to have merit in documenting further changes in water volume, solids load, and stream temperature after the residential development is completely built out.

Critical to obtaining useful conclusions for this study was the ability to document that downstream loads were significantly greater than upstream loads before any BMPs were in place. Results from the Wilcoxon signed ranks test, used to find differences between paired data sets, revealed that downstream loads were significantly greater than upstream loads at the 90-percent confidence level. Therefore, the study area was an important contributor of total and suspended solids to Brewery Creek. However, previous studies indicate that streambank slumping could be an additional input to the solids load of the stream in addition to inputs related to construction activities (Allen and Gray, 1984).

Summary statistics for solids load at the downstream and upstream gages during each phase of the study are listed in table 3. Mean volume, total solids load, and total suspended solids load are greater at the downstream site than the upstream site for each phase of construction. However, examination of downstream and upstream volumes and loads revealed a highly skewed distribution. Large rain events can skew the distribution of volume and solids load. One such event occurred in August 2001 when over 9.5 in of rain was recorded at the downstream station within 48 hours. This type of event is atypical and should be given less weight statistically. Median rather than mean values were used during statistical analyses because the median is a more appropriate representation of the population center in highly skewed data sets than the mean. (Ott and Longnecker, 2001).

The difference between downstream and upstream loads was computed for total and suspended solids for the preconstruction, land-disturbance, and home-construction phases. Changes in the magnitude of the differences are believed to be a result of activity in the study area. The erosion-control and stormwater BMPs used at the construction site were effective at limiting the amount of solids load entering Brewery Creek (fig. 10). Each bar represents the median of all differences between the downstream and upstream constituent loads for the preconstruction, land-disturbance, and home-construction storm-runoff periods. Differences in median total solids loads decreased by 52 percent between the preconstruction and land-disturbance phases and 22 percent between the land-disturbance and home-construction phases (fig. 10). Similarly, downstream-upstream differences in suspended-solids loads decreased 72 and 37 percent, respectively (fig. 10).

However, examination of the median value fails to explain the variability of the data. Most of the coefficients of variation in table 3 have a value greater than 1, indicating substantial variability in solids load. The Wilcoxon rank sum test (Ott and Longnecker, 2001) was used to describe the variability of the data and to ultimately determine whether the contribution of loads significantly increased or decreased from one phase of construction to the next. The null hypothesis states there is no change in the magnitude of the difference between the upstream and downstream solids load from one phase of construction to the next. The alternative hypothesis suggests there is a significant change in the magnitude of the difference between the downstream and upstream solids load from one phase of construction to the next and this change is related to BMP effectiveness. Results from the test, at the 90-percent confidence level, failed to reject the null hypothesis. The data provide insufficient evidence to report an increase or decrease in solids load from preconstruction levels. Because the test did not indicate a significant increase in solids load, one could imply the BMP systems implemented before and during the land-disturbance and home-construction phases are at least somewhat effective at limiting the amount of solids entrained in runoff from reaching Brewery Creek. This limitation is supported by the reduction in the magnitude of differences between downstream and upstream total solids and total suspended solids load from one phase of construction to the next (fig. 10).

### Runoff Volume

Summary statistics for event volumes are also detailed in table 3. Similar to solids loads, the median of all differences between the upstream and downstream volumes were determined for each phase of the study. A 60-percent reduction in median runoff volumes from the preconstruction to the land-disturbance phase is illustrated in figure 11; results from the Wilcoxon rank sum test shows this difference to be significant at the 90-percent confidence level.

Median runoff volumes appeared to increase slightly between the land-disturbance to the home-construction phases (fig. 11); however, statistical tests indicated no significant difference in runoff volume between these two phases. The apparent increase could be due, in part, to the failure of a runoff infiltration pond during the home-construction phase. Overall, the BMPs utilized within the study area were able to reduce the amount of stormwater runoff entering Brewery Creek.

### Stream Temperature

The temperature of urban streams is often affected directly by urban runoff. For example, Galli (1990) demonstrated an increase in base flow water temperature of 0.14°C for every 1-percent increase in watershed imperviousness. Although the Brewery Creek Watershed contains only 3 percent urban land, the stream flows through an urban environment for approximately 0.5 mi before its confluence with Black Earth Creek, a Class I trout stream. Certain species of fish, such as trout, require relatively low daily mean temperatures of less than 22°C (Lyons, 1996) for survival and are particularly sensitive to temperature fluctuations. As urbanization continues to spread throughout the basin, mitigation of thermal impacts caused by increases in impervious surfaces will be increasingly important.

A summary of daily mean stream temperatures measured during sampled events in each phase of construction is given in table 4. Downstream temperatures were statistically higher than upstream temperatures, most likely because lack of overhead tree canopy in the area between the upstream and downstream stations subjects the stream to direct solar heating. To determine whether the downstream temperatures increased as a result of activity within the study area, a one-way analysis of variance (ANOVA) (Ott and Longnecker, 2001) test was used to identify differences between the means of independent samples. Test results showed no significant increases in stream temperature as a result of activity within the study area (at the 90-percent confidence level).

## Ecologic Response

Ecologic response in terms of macroinvertebrate communities, fish communities, and habitat to construction of the residential subdivision is discussed in the following sections below.

### Macroinvertebrate and Fish Communities

A total of 10 macroinvertebrate samples were collected for this study from October 1999 to October 2002. HBI values (fig. 12) ranged from 4.19 (very good water quality) to 4.76 (good water quality). No significant differences (P=0.05) were detected either between sampling sites or before and after development. Results indicate that dissolved-oxygen concentrations were consistently sufficient to support a diverse macroinvertebrate community and that organic contamination was not appreciable throughout the 3-year study period. Interpretation of the empirical results, relative to Hilsenhoff's scale, was that the degree of organic pollution ranged from "possible slight to some" during the study period.

Fish communities were sampled six times for this study. A total of 10 species were collected and identified at least once. The list includes brown trout, creekchub, fathead minnow, golden shiner, white sucker, yellow bullhead, black bullhead, brook stickleback, green sunfish, and bluegill. The species in bold are considered tolerant to environmental degradation. No intolerant species were found. The proportion of tolerant individuals ranged from 54 percent to 78 percent (fig. 13) and was the primary reason IBI scores remained low throughout the entire study. IBI scores of 10 or 20 (fig. 13) were both within the "poor" range. Brown trout were relatively abundant in Brewery Creek with numbers ranging from 22 to 46 (fig. 14) and percentages ranging from 16 percent to 46 percent (fig. 13). Brown trout size structure was variable during the study, with greater numbers of juvenile individuals from 1999 through 2001 and greater numbers of adults in 2002 and 2003. Trout of legal keeping size (greater than 9 in. long) ranged from 1 out of 37 in 2001 to 15 out of 26 in 2002 and 15 out of 44 in 2003 (fig. 14). The largest brown trout was 15.75 in. long.

Low IBI scores reflected a combined absence of intolerant species and a general lack of coldwater indicators (with the exception of brown trout) and presence of numerous tolerant species. Although brown trout size structure did change during the study, the overall fish-community structure remained the same.

Differences in the macroinvertebrate and fish sample results are related to different metric objectives. The HBI is based on macroinvertebrate tolerances to organic pollution, whereas the coldwater IBI is based on fish tolerances to a wide variety of environmental factors including temperature and physical habitat. The combination of these results indicate that organic-contaminant loading is not a limiting factor in Brewery Creek but that overall habitat is in poor condition.

Beginning in the mid-1980s, Brewery Creek became the focus of water-quality evaluation as part of the Black Earth Creek Priority Watershed Project. Historical water-quality, macroinvertebrate, and fish-community data indicated that Brewery Creek was appreciably impaired. Best management practices implemented as part of the Black Earth Creek Priority Watershed Project did not affect total-phosphorus or suspended-sediment concentrations over time, but ammonia concentrations did decline (Graczyk and others, 2003). This result suggests that organic loading declined. Such a decline is also reflected by a trend (R2 = 0.58) of improved HBI scores (fig. 15). Although fish-community data from Brewery Creek are limited, surveys done before 1990 indicate that the stream was degraded (fig. 13). No trout were found during surveys in 1979 or 1989, and IBI scores were 0 or "very poor". Beginning in 1999, substantial brown trout numbers were found in every survey, and coldwater IBI scores improved slightly. Although the overall fish community is still considered unbalanced, the recent fish-shocking surveys are consistent with macroinvertebrate collections and indicate improved water quality and habitat conditions in Brewery Creek.

The improved water-quality and habitat conditions in Brewery Creek are beneficial for managing Black Earth Creek trout fisheries. Not only have organic loads declined in Brewery Creek, the small tributary also provides habitat for migrating brown trout and forage populations. During the 3-year study, fisheries in Brewery Creek were not affected as a result of the new subdivision development.

### Habitat

Preconstruction results from sampling locations H1 and H2 yielded Habitat Index Scores of 40 and 45, respectively (table 5), which correspond to "fair" in the qualitative assessment. Scores from the postconstruction evaluation were 40 and 35 at H1 and H2, respectively, but the rating for H2 was still "fair." The 10-point change at H2 was due to 5-point changes in two of the assessment metrics, width-to-depth ratio and riffle-to-riffle ratio. The change in width-to-depth ratio resulted from placement of a large double-celled culvert within the habitat station (fig. 16).

Mean stream width and depth also were computed for each habitat station (table 5). The preconstruction mean stream widths indicate a second-order stream (Strahler, 1957). Postconstruction mean stream widths indicate a substantial increase for both stations. The change was greater at H2, where the difference of 3.3 feet amounted to a 53-percent increase. Changes in mean stream depth from preconstruction to postconstruction were, in contrast, insubstantial and opposite for the two stations, H1 deepening by 0.13 feet and H2 becoming shallower by 0.13 feet. The changes at H2 were, again, attributable to the placement of the large culvert.

[ft, feet]

Site Mean stream width (ft) Mean stream depth (ft) Width/depth ratio Habitat score
H1 Preconstruction 9.3 0.9 10.3 40/FAIR
H2 Preconstruction 6.3 1.0 6.3 45/FAIR
H1 Preconstruction 10.2 1.0 10.2 40/FAIR
H2 Preconstruction 9.6 0.9 10.7 35/FAIR

The mean depth of fines over the coarse sand and silt substrate at both stations ranged from 0.26 to 0.35 feet (table 6). At H1, a slight postconstruction increase in mean depth of fines was noted. At H2, however, the depth of 4.86 feet was nearly double the preconstruction value.

[all depths in feet]

Site Mean Maximum Median
Preconstruction H1 0.35 1.18 0.33
Preconstruction H2 0.26 0.92 0.23
Preconstruction H1 0.44 1.18 0.43
Preconstruction H2 0.49 1.18 0.39
Difference H1 0.09 0 0.10
Difference H2 0.23 0.26 0.16

The exact cause for the increase is difficult to determine. Sediment influx as a result of construction activity of the St. Francis development is one possible scenario; however, results of water-quality sampling during storm events did not substantiate this scenario. Other activities or occurrences within the study area may have contributed to an increase in the depth of fines, including removal of the weir structure during spring 2003, contribution of fines from upstream sources, and failing streambanks. Data from additional habitat sites would be needed to identify other potential sources of sediment to Brewery Creek.

## Geomorphic Response

Results of the bank-pin surveys indicated failing banks at most sites. Bank pins in cross sections 1–8 (straightened reach) averaged 43.6 mm of deposition, whereas bank pins at cross sections 9–14 (meandering reach) averaged 26.4 mm of deposition at the conclusion of the study period. Bank material on Brewery Creek is primarily cohesive alluvial silt loam, making exposed banks susceptible to erosion. Although measurements indicated deposition, this was due to bank failure rather than sediment deposition on the banks. Streambank erosion processes are classified into two basic groups: gravitational or mechanical failures and tractive-force failures (O'Neill and Kuhns, 1994). The failing banks on Brewery Creek were indicative of gravitational failure, given low flows and the fine-grained cohesive soils. An example of streambank failure on Brewery Creek is shown in figure 17.

Results of the Rosgen Stream Classification showed only minor variations from 2001 and 2003 classifications (table 7). Bankfull widths and average depths decreased, whereas the width depth (w/d) ratios and entrenchment ratios increased. These changes were due in part to the placement of two large culverts (fig. 16) directly upstream from cross-section 8 and lower annual average precipitation in 2003. At cross section 1, the stream type went from a "G" to an "F" (Rosgen, 1996) with the dominant bed material changing from gravel to silt-clay. This difference can be attributed to removal of the V-notch weir at the upstream water-quality station in May 2003. During removal of the weir, soft sediment that had been deposited upstream from the structure was flushed and allowed to travel downstream. Cross section 9, in the uppermost part of the meandering section, did not change in stream type, but an increase in size of bed material from silt-clay to more gravel was noted.

Results of the Wolman Pebble Count Procedure indicated an increase in the cumulative percentage of fine-grained sediment at all transect survey locations. For example, differences in size-class distribution at cross section 13 can be seen in figure 18.

In summary, results of the fluvial geomorphology classifications and analyses do not indicate that the St. Francis Development on Brewery Creek contributed to changes in stream characteristics. Stream hydrology may have been altered slightly by removal of the weir and placement of the culverts as part of the road construction, but it is difficult to quantify the effects of these actions, if they can be quantified at all. No significant changes were detected after analyzing yearly stream-survey data.

Erosion of the streambanks was the primary source of increased fine-grained sediment noted during annual pebble counts. Based on flow regimes and overall decreases in sediment yield from the construction site, the percentage of fine-grained sediment should have decreased corresponding to flows and constituent loading. In the loess area of the midwestern United States, however, bank material has been reported to contribute as much as 80 percent of the total sediment eroded from incised channels (Simon and others,1996). Analysis with Rosgen's stream classification indicated that channels were entrenched to slightly entrenched, making them more susceptible to erosion processes.

In addition to bank erosion, removal of the V-notch at the upstream water-quality station in the spring of 2003 likely contributed to an increase in fine-grained sediments. The weir was in the upstream segment of the study area (channelized reach) and released sediment downstream when it was removed. Data collected in 2001 and 2002 were very similar in results, whereas 2003 data indicated a significant increase in fines at all measured cross sections.

## Summary and Conclusions

The U.S. Geological Survey (USGS), in cooperation with the Dane County Land Conservation Department (LCD) and the Wisconsin Department of Natural Resources (WDNR), conducted a multidisciplinary study incorporating streamflow, water-quality sampling, and physical, ecological, and geomorphic metrics to assess instream effects from construction of a residential subdivision on Brewery Creek, Dane County, Wis. An upstream/downstream (above and below) approach was used to isolate any changes caused by the study area over a period of 3 years (2001–03).

Collectively, the stormwater-management and erosion-control BMPs used at the St. Francis residential subdivision provided sufficient protection against degradation to Brewery Creek. Additionally, proper implementation and maintenance of the erosion-control and stormwater-management plan were critical components to reducing stormwater runoff. Results from this project will serve as an example for Dane County developers and builders of how to meet stormwater standards detailed in the Dane County Ordinance.

Erosion and stormwater-management controls implemented within the study area were effective at controlling runoff and solids transport during construction activity. Downstream event volumes, loads, and temperature were significantly greater than upstream volumes, loads, and temperature during three phases of construction: preconstruction, land disturbance, and home construction. The effectiveness of stormwater-management and erosion-control BMP systems was measured by evaluating the change in magnitude of differences between the downstream and upstream stations from one phase of construction activity to the next. The median difference between downstream and upstream storm volumes decreased (60 percent) from the preconstruction phase to the land-disturbance phase and slightly increased (9 percent) from the land-disturbance phase to the home-construction phase. The median differences for total and suspended solids load indicated a similar trend from preconstruction to land-disturbance phases with decreases of 52 and 72 percent, respectively. Both total and suspended solids load continued to decrease in the transition, from land-disturbance to home-construction phases; by 22 and 37 percent, respectively. Extreme data variability hampered statistical interpretation. Additional storm volume and load data could reduce variability and improve the statistical significance when determining an increase or decrease in volume or load from the study area.

Although daily mean stream temperature at the downstream monitoring station was consistently higher than at the upstream monitoring station during each phase, there was no statistical evidence to suggest an increase in stream temperature as a result of activity within the study area. Stream temperatures were most likely affected by direct solar heating because of a lack of overhead tree canopy between the downstream and upstream stations on Brewery Creek. Tree canopy was not altered during construction activity in the study area and was not considered part of the storm-runoff BMP system.

Ecologic indices for macroinvertebrate and fish communities indicate there were no negative effects to water quality and fisheries in Brewery Creek as a result of activity within the St. Francis subdivision. Macroinvertebrate sampling results (HBI value) on Brewery Creek ranged from "very good" to "good" water quality with no significant differences during any phase of construction activity. Results for fish-community composition, however, fell within the "poor" range (IBI value) during each year of testing. A general absence of intolerant species, with the exception of brown trout, reflects the low IBI values. The combination of these results suggests that organic loading is not a limiting factor in Brewery Creek but that overall fish habitat is in poor condition.

Habitat measurements did not change significantly from preconstruction to postconstruction phases. Although installation of a double-celled culvert in Brewery Creek most likely altered the width-to-depth ratio in that reach, the overall habitat rating remained "fair". Installation of the culvert may also have caused changes in mean stream width and depth. These changes were a result of modifications to the stream itself and do not reflect changes caused by surface runoff because of activities in the study area.

Fluvial geomorphology classifications, including channel cross sections, bed- and bank-erosion surveys, and pebble counts did not indicate that stream geomorphic characteristics were altered by home-construction activity in the study area. Increases in fine-grained sediment at various cross sections were attributed to instream erosion processes, such as bank slumping, rather than increases in sediment delivery from the nearby construction site. This result was further substantiated by the reduction of storm runoff from the construction site during each phase of the study. Additional sediment was introduced to the stream by way of removal of the V-notch weir at the upstream monitoring station in spring 2003.

## Acknowledgments

The authors would like to thank Jesse Pritts of the U.S. Environmental Protection Agency for assisting in development of the study; Daniel Heffron, property owner and developer, for graciously allowing us access to the site and for agreeing to implement the stormwater management plan; Ron Steiner, project engineer, for coordinating and developing the stormwater management plan; Aicardo Roa of Dane County Land Conservation Department for his assistance on the stormwater plan and site assessment; Dave Owens of the U.S. Geological Survey for assisting with the instrumentation; Faith Fitzpatrick of the U.S. Geological Survey and Barb Lensch of the Natural Resources Conservation Service for assisting with aspects of the study related to stream geomorphology; and the staff of the Dane County Land Conservation Department for countless hours of support. Without their cooperation and effort, this study would not have been possible.

## References

Allen, H.E., and Gray, J.R., 1984; Runoff, sediment transport, and water-quality in a Northern Illinois agricultural watershed before urban development, 1979–81: U.S. Geological Survey Water-Resources Investigation Report 82–4073, 55 p.

Dane County Code of Ordinances [2001], Chapter 14, subchapter II, Erosion control and stormwater management: Madison, Wis., Dane County Board of Supervisors, 15 p.

Fago, Don, 1992, Distribution and relative abundance of fishes in Wisconsin: Wisconsin Department of Natural Resources Technical Bulletin 175, 378 p.

Galli, J., 1990a, Thermal impacts associated with urbanization and stormwater management best management practices; appendix C, Water temperature and freshwater stream biota; an overview: Washington D.C., Metropolitan Washington Council of Governments.

Glocker, C.L., and Patzer, R.A., 1978, Soil survey of Dane County: U.S. Department of Agriculture Soil Conservation Service.

Graczyk, D.J., Walker, J.F., Horwatich, J.A., and Bannerman, R.T., 2003, Effects of best management practices in the Black Earth Creek Priority Watershed, Wisconsin, 1984–98: U. S. Geological Survey Water-Resources Investigation Report 03–4163, 24 p.

Harrelson, C.C., Rawlins, C.L., and Potyondy, J.P., 1994, Stream channel reference sites—An illustrated guide to field technique: U.S. Department of Agriculture Forest Service General Technical Report RM-245, 61 p.

Hilsenhoff, W.L., 1987, An improved biotic index of organic stream pollution: The Great Lakes Entomologist, v. 20, no. 1, p. 31–39.

Lyons, John, Wang, Lizhu, and T.D. Simonson, 1996, Development and validation of an index of biotic integrity for coldwater streams in Wisconsin: North American Journal of Fisheries Management, v. 16, p. 241–256.

O'Neill, M.P., and Kuhns, M.R., 1994, Stream bank erosion and flushing flows: Stream Systems Technology Center Stream Notes, July 1994, p. 3–6.

Ott, L.R., and Longnecker, M., 2001, An introduction to statistical methods and data analysis, fifth edition: Wadsworth Group, Pacific Grove, Calif., [variously paginated].

Porterfield, George, 1972, Computation of fluvial-sediment discharge: U.S. Geological Survey Techniques of Water-Resources Investigations, book 3, chap. C2, 66 p.

Rosgen, Dave, 1996, Applied river morphology: Pagosa Springs, Colo., Wildland Hydrology, [variously paginated].

Simonson, T.D., Lyons, J., and Kanehl, P.D., 1994, Guidelines for evaluating fish habitat in Wisconsin streams: U.S. Department of Agriculture, General Technical Report NC-164, 36 p.

Strahler, A.N., 1957, Quantitative analysis of watershed geomorphology: American Geophysical Union Transactions, v. 38, p. 913–920.

U.S. Environmental Protection Agency, 1992, Environmental impacts of stormwater discharges, a national profile: Washington, D.C., Office of Water, EPA 841–R–92–001.

U.S. Environmental Protection Agency, 1997, Urbanization and streams—Studies of hydrologic impacts: Washington, D.C., Office of Water, EPA 841–R–97-009.

U.S. Environmental Protection Agency, 2001, Techniques for tracking, evaluating, and reporting the implementation of nonpoint source control measures—III. Urban: Washington, D.C., Office of Water.

U.S. Environmental Protection Agency, 2002. National water-quality inventory, 2000 report: Washington, D.C., Office of Water, EPA 841-R-02-001.

Ward, J.R. and Harr, C.A., eds., 1990, Methods for collection and processing of surface-water and bed-material samples for physical and chemical analyses: U.S. Geological Survey Open-File Report 90–140, 71 p.

Wisconsin Nonpoint Source Water Pollution Abatement Program, 1989, A plan for control of nonpoint sources and related resource management in the Black Earth Creek Priority Watershed: Wisconsin Department of Natural Resources Publication WR-218-89, [variously paginated].

## Appendix

[-, storm duration does not allow for intensity computation; in/hr, inches per hour]

Start date/time   End date/time Total rain (inches) 5-minute intensity (in/hr) 10-minute intensity (in/hr) 15-minute intensity (in/hr) 30-minute intensity (in/hr) 60-minute intensity (in/hr)   Erosivity index
10/01/99 21:05 10/02/99 04:30 .36 .24 .18 .16 .12 .10 .27
10/03/99 12:15 10/03/99 18:00 .21 .12 .12 .12 .10 .09 .13
10/16/99 01:20 10/16/99 07:10 .25 .48 .42 .36 .26 .18 .44
11/10/99 14:45 11/10/99 18:30 .80 1.92 1.86 1.84 1.10 .72 7.83
11/23/99 01:45 11/23/99 12:40 .99 .84 .66 .52 .36 .21 2.52
12/03/99 10:55 12/03/99 18:40 .31 .12 .12 .12 .12 .09 .23
12/04/99 15:00 12/04/99 23:30 .12 .12 .06 .08 .06 .05 .04
12/09/99 11:00 12/09/99 16:40 .15 .12 .12 .08 .08 .06 .07
01/09/00 20:30 01/10/00 07:55 .13 .12 .06 .08 .06 .04 .05
02/15/00 12:35 02/15/00 13:50 .16 .24 .24 .20 .18 .14 .19
02/24/00 02:40 02/24/00 10:30 .31 .12 .12 .12 .10 .08 .19
02/25/00 18:10 02/26/00 01:25 .66 1.32 .96 .72 .42 .27 1.94
03/08/00 10:40 03/08/00 17:50 .20 .72 .60 .52 .30 .16 .46
03/09/00 03:25 03/09/00 08:45 .13 .12 .12 .12 .10 .07 .08
03/15/00 12:35 03/15/00 15:30 .13 .12 .12 .12 .08 .06 .06
03/19/00 13:05 03/20/00 05:35 .34 .12 .12 .08 .08 .06 .17
03/26/00 16:15 03/26/00 18:35 .11 .24 .24 .20 .14 .08 .10
04/08/00 13:20 04/08/00 15:15 .28 .36 .36 .36 .32 .25 .64
04/19/00 07:35 04/21/00 00:40 1.67 .96 .72 .64 .42 .38 5.03
04/23/00 00:15 04/23/00 09:30 .46 .24 .18 .16 .14 .14 .41
05/08/00 08:25 05/08/00 10:20 .13 .48 .42 .36 .24 .12 .23
05/11/00 22:35 05/12/00 09:20 .14 .24 .24 .20 .10 .07 .09
05/17/00 12:50 05/20/00 07:35 5.04 4.68 4.38 3.80 2.42 1.24 101.33
05/26/00 22:30 06/01/00 23:30 5.071 11.52 6.60 4.44 2.24 1.36 103.42
06/04/00 06:25 06/05/00 06:55 .88 .36 .24 .20 .20 .19 1.13
06/12/00 06:00 06/12/00 11:10 .20 .36 .36 .32 .26 .15 .36
06/13/00 15:15 06/14/00 02:40 2.71 3.00 2.88 2.44 1.52 .87 37.54
06/16/00 03:20 06/16/00 05:40 .11 .12 .12 .12 .10 .06 .07
06/20/00 04:35 06/20/00 16:00 .70 .96 .90 .76 .50 .31 2.58
06/24/00 14:00 06/24/00 18:50 .29 .24 .18 .16 .14 .13 .25
06/26/00 08:50 06/26/00 10:00 .11 .36 .30 .20 .12 .10 .09
06/28/00 06:20 06/28/00 10:35 .18 .24 .18 .16 .12 .09 .13
07/02/00 19:35 07/02/00 21:15 .99 3.48 3.12 2.68 1.64 .90 15.65
07/09/00 06:20 07/09/00 10:15 .42 .48 .36 .32 .24 .22 .69
07/10/00 03:40 07/10/00 08:45 .73 .84 .72 .68 .44 .30 2.34
07/20/00 17:00 07/20/00 17:15 .14 .96 .78 .56 - - -
07/28/00 17:35 07/28/00 22:50 .20 .60 .48 .40 .30 .18 .44
07/30/00 16:05 07/30/00 16:20 .14 .96 .78 .56 - - -
08/05/00 11:15 08/05/00 22:10 1.81 3.12 3.00 2.80 2.06 1.39 34.29
08/12/00 19:50 08/12/00 21:15 .13 .12 .12 .12 .12 .11 .10
08/13/00 06:30 08/13/00 09:35 .14 .36 .30 .28 .18 .11 .17
08/16/00 20:20 08/17/00 09:05 1.00 1.20 1.02 .96 .72 .50 5.48
08/26/00 07:30 08/26/00 10:05 1.14 2.88 2.34 1.84 1.28 .91 12.99
09/03/00 07:05 09/03/00 08:30 .28 1.08 .72 .56 .36 .24 .76
09/11/00 07:40 09/11/00 22:05 1.081 2.04 1.80 1.48 .90 .77 8.59
09/14/00 00:45 09/14/00 06:40 .22 .36 .30 .28 .20 .10 .29
09/19/00 17:00 09/19/00 23:55 .54 .84 .54 .40 .28 .22 1.04
09/22/00 09:20 09/22/00 23:30 .64 .36 .30 .28 .18 .17 .75
10/03/00 19:30 10/03/00 23:45 .17 .24 .18 .16 .12 .10 .13
10/05/00 16:10 10/05/00 20:00 .16 .24 .18 .16 .12 .07 .12
10/23/00 08:20 10/23/00 14:25 .29 .24 .18 .16 .12 .10 .22
11/06/00 13:00 11/07/00 00:30 .89 .48 .36 .32 .24 .18 1.41
11/09/00 09:50 11/09/00 16:30 .15 .48 .24 .16 .08 .06 .08
11/24/00 11:25 11/24/00 14:15 .25 .24 .18 .16 .14 .11 .22
12/15/00 20:10 12/16/00 06:45 .35 .24 .18 .16 .12 .08 .26
12/18/00 09:25 12/18/00 22:20 .29 .12 .12 .08 .06 .06 .11
12/28/00 19:45 12/29/00 19:05 .22 .12 .06 .08 .04 .03 .05
01/14/01 01:45 01/14/01 09:55 .24 .12 .12 .12 .12 .09 .18
01/29/01 11:30 01/29/01 23:05 .86 .36 .30 .28 .22 .18 1.21
02/07/01 21:30 02/08/01 04:15 .24 .36 .36 .28 .26 .18 .43
02/08/01 15:25 02/09/01 17:55 1.34 .48 .42 .36 .32 .29 2.89
02/24/01 06:20 02/24/01 20:10 .23 .12 .12 .08 .08 .07 .11
03/12/01 07:45 03/12/01 11:30 .14 .12 .12 .08 .08 .07 .07
03/31/01 14:15 03/31/01 17:15 .27 .24 .18 .16 .14 .13 .24
04/05/01 13:15 04/05/01 17:35 .31 .72 .42 .36 .28 .15 .64
04/08/01 23:20 04/09/01 04:35 .76 .36 .36 .32 .28 .24 1.46
04/10/01 22:50 04/11/01 15:20 .73 1.20 .78 .68 .50 .31 2.82
04/20/01 02:15 04/20/01 06:10 .48 .36 .30 .28 .24 .20 .77
04/21/01 00:10 04/21/01 04:55 .44 1.44 .72 .56 .38 .23 1.37

1 storm defined using 12-hour interval

[-, storm duration does not allow for intensity computation; in/hr, inches per hour]

Start date/time   End date/time Total rain (inches) 5-minute intensity (in/hr) 10-minute intensity (in/hr) 15-minute intensity (in/hr) 30-minute intensity (in/hr) 60-minute intensity (in/hr)   Erosivity index
05/01/01 04:05 05/01/01 07:10 .15 .36 .30 .28 .22 .13 .23
05/03/01 18:30 05/04/01 08:35 .37 .36 .30 .28 .18 .12 .43
05/06/01 19:35 05/07/01 10:50 .46 .48 .36 .36 .26 .21 .83
05/10/01 00:55 05/10/01 03:05 .15 .36 .30 .28 .20 .13 .21
05/10/01 19:45 05/11/01 05:15 .60 1.68 1.50 1.16 .70 .41 3.44
05/20/01 23:55 05/21/01 12:20 2.02 1.92 1.74 1.52 1.04 .71 17.33
05/22/01 11:20 05/23/01 16:30 .54 .48 .48 .32 .18 .11 .65
05/31/01 15:20 06/01/01 07:30 .42 .24 .18 .12 .10 .08 .26
06/01/01 16:55 06/01/01 20:20 .45 1.32 .78 .60 .40 .29 1.36
06/05/01 01:30 06/05/01 13:00 .73 1.32 1.02 .80 .54 .41 2.97
06/09/01 21:35 06/10/01 05:30 .28 .72 .54 .44 .26 .13 .56
06/11/01 22:00 06/12/01 05:30 2.07 2.40 2.22 2.04 1.74 1.31 33.12
06/14/01 11:40 06/14/01 22:55 .14 .72 .48 .32 .16 .08 .17
06/15/01 05:40 06/15/01 10:45 .16 .24 .18 .16 .08 .08 .08
06/18/01 03:00 06/18/01 11:55 .46 .48 .42 .36 .22 .19 .69
06/21/01 18:50 06/22/01 00:15 .33 .48 .42 .44 .34 .26 .81
07/17/01 08:20 07/17/01 13:45 .55 3.00 1.86 1.40 .80 .52 3.94
07/18/01 07:15 07/18/01 08:30 .98 3.72 3.24 2.76 1.74 .96 16.70
07/24/01 20:25 07/25/01 02:00 .24 .72 .48 .32 .16 .09 .26
08/01/01 18:00 08/02/01 08:00 9.56 5.40 5.10 4.76 3.84 2.74 235.11
08/09/01 17:10 08/09/01 17:25 .11 1.08 .60 .44 - - -
08/15/01 15:25 08/16/01 07:30 .50 .24 .24 .20 .14 .09 .43
08/22/01 06:40 08/22/01 13:05 .20 .24 .18 .20 .16 .08 .21
08/24/01 19:10 08/26/01 01:40 1.31 2.28 1.98 1.44 1.08 .84 12.09
08/27/01 04:20 08/27/01 06:30 .16 .48 .30 .24 .14 .09 .15
09/06/01 14:15 09/07/01 03:30 .14 .72 .42 .28 .14 .07 .14
09/07/01 12:20 09/07/01 12:40 .46 2.52 2.34 1.76 - - -
09/07/01 19:05 09/08/01 03:00 1.28 2.52 2.22 2.04 1.48 .81 17.45
09/09/01 09:00 09/09/01 19:50 .71 .36 .24 .24 .18 .16 .82
09/17/01 04:45 09/17/01 12:15 .57 .36 .30 .28 .24 .22 .90
09/18/01 23:45 09/19/01 14:35 1.06 .36 .30 .28 .22 .22 1.52
09/20/01 19:50 09/21/01 00:30 .24 .24 .24 .24 .24 .21 .40
09/22/01 22:30 09/23/01 16:40 1.73 1.08 .96 .88 .84 .63 10.80
10/09/01 23:35 10/10/01 12:55 .23 .24 .18 .16 .14 .08 .21
10/13/01 09:15 10/13/01 20:40 .20 .36 .24 .16 .08 .07 .10
10/22/01 14:40 10/22/01 22:05 1.17 .96 .84 .76 .60 .41 5.06
10/24/01 07:25 10/24/01 18:35 .25 .24 .24 .20 .18 .14 .29
11/13/01 02:50 11/13/01 10:05 .24 .36 .24 .20 .14 .12 .22
11/18/01 18:20 11/19/01 02:15 .25 .24 .12 .12 .10 .07 .16
11/23/01 18:00 11/24/01 11:00 .82 .60 .54 .48 .30 .19 1.62
11/26/01 17:20 11/27/01 02:45 .25 .36 .24 .20 .12 .08 .19
11/30/01 02:10 11/30/01 07:05 .25 .24 .18 .16 .14 .10 .22
12/05/01 03:05 12/05/01 07:55 .17 .24 .18 .20 .12 .10 .13
12/12/01 17:30 12/13/01 04:05 .58 .48 .42 .36 .28 .23 1.10
12/22/01 07:55 12/22/01 18:00 .23 .36 .30 .24 .18 .10 .27
02/01/02 10:05 02/01/02 13:20 .31 .24 .18 .16 .16 .13 .31
02/09/02 22:45 02/10/02 13:20 .50 .12 .12 .12 .12 .09 .37
02/18/02 22:10 02/20/02 19:55 1.68 .24 .24 .20 .18 .16 1.89
03/05/02 12:15 03/05/02 15:20 .20 .24 .18 .20 .16 .13 .20
03/07/02 22:00 03/09/02 09:45 .461 .84 .54 .44 .24 .14 .80
03/19/02 13:50 03/19/02 22:10 .20 .12 .12 .12 .10 .08 .12

1 storm defined using 12-hour interval

[-, storm duration does not allow for intensity computation; in/hr, inches per hour]

Start date/time   End date/time Total rain (inches) 5-minute intensity (in/hr) 10-minute intensity (in/hr) 15-minute intensity (in/hr) 30-minute intensity (in/hr) 60-minute intensity (in/hr)   Erosivity index
04/02/02 08:25 04/02/02 18:30 .25 .12 .12 .12 .10 .09 .15
04/07/02 03:40 04/09/02 06:40 1.361 .36 .36 .32 .28 .14 2.40
04/12/02 00:25 04/12/02 02:20 .16 .24 .24 .20 .18 .11 .19
04/14/02 18:10 04/14/02 19:10 .19 1.08 .84 .64 .32 .19 .51
04/18/02 16:00 04/18/02 22:25 .66 1.44 1.32 1.08 .84 .46 4.74
04/21/02 03:15 04/22/02 03:00 .32 .36 .24 .24 .16 .08 .32
04/24/02 12:55 04/24/02 15:30 .39 .72 .66 .64 .58 .36 1.79
04/27/02 11:05 04/28/02 06:15 .70 .36 .30 .28 .18 .12 .79
05/01/02 12:20 05/02/02 00:20 .56 .36 .30 .28 .22 .15 .79
05/06/02 22:15 05/06/02 23:15 .31 1.32 1.14 1.00 .60 .31 1.67
05/08/02 22:20 05/09/02 07:45 .78 1.44 .90 .64 .64 .50 3.90
05/11/02 10:50 05/12/02 10:30 .83 .84 .42 .32 .28 .22 1.60
05/25/02 05:05 05/25/02 13:55 .76 .36 .36 .32 .30 .24 1.50
05/28/02 20:25 05/29/02 05:35 .30 .60 .54 .44 .36 .24 .79
06/02/02 16:30 06/03/02 10:15 .24 .24 .18 .16 .12 .09 .18
06/03/02 22:45 06/04/02 11:30 .36 .36 .36 .32 .24 .16 .56
06/10/02 18:55 06/10/02 20:50 .14 .24 .24 .24 .16 .11 .15
06/26/02 04:05 06/26/02 10:10 .39 1.20 .84 .68 .38 .21 1.10
07/08/02 12:05 07/08/02 12:45 .56 1.56 1.50 1.48 1.06 - 5.56
07/08/02 20:35 07/08/02 20:50 .10 .60 .54 .40 - - -
07/20/02 15:20 07/20/02 15:55 .55 2.28 2.16 1.68 1.08 - 5.75
07/22/02 00:35 07/22/02 07:25 1.16 3.48 2.46 2.12 1.28 .70 13.21
07/27/02 05:30 07/27/02 10:30 .21 .24 .18 .16 .10 .08 .13
07/27/02 23:05 07/27/02 23:15 .12 1.20 .72 - - - -
08/04/02 02:25 08/04/02 10:35 .73 2.28 2.04 1.76 .98 .50 6.34
08/11/02 16:35 08/11/02 17:55 1.74 3.96 3.30 2.96 2.02 1.53 35.90
08/12/02 20:15 08/13/02 05:10 .43 .60 .42 .36 0.26 .16 .76
08/17/02 06:50 08/17/02 07:10 .32 1.56 1.44 1.12 - - -
08/21/02 18:00 08/22/02 13:05 .931 2.28 2.10 1.60 .88 .45 6.62
09/02/02 03:45 09/02/02 08:15 1.05 1.56 1.38 1.20 .90 .57 7.73
09/10/02 12:15 09/10/02 13:25 .16 .24 .24 .20 .18 .14 .19
09/19/02 00:55 09/19/02 05:05 .35 .72 .42 .36 .22 .17 .54
09/19/02 12:45 09/19/02 20:35 .17 .72 .48 .48 .26 .14 .34
09/20/02 04:00 09/20/02 18:15 .28 .96 .72 .52 .28 .14 .57
09/28/02 21:45 0 9/29/02 05:35 .70 1.45 1.17 .93 .71 .36 3.99

1 storm defined using 12-hour interval

[mg/L, milligrams per liter; RPD, relative percent difference; NA, not applicable; <, less than; %, percent]

Date/time Sample type Total solids (mg/L) Total solids (RPD) Suspended solids (mg/L) Suspended solids (RPD)
UPSTREAM
08/17/99 16:45 Blank 66 NA 48 NA
08/17/99 17:00 Blank <7 NA <5 NA
04/17/01 09:40 Blank <20 NA <2 NA
07/09/02 09:20 Blank <50 NA <2 NA
08/21/02 13:05 Replicate 454 4% 18 0%
08/21/02 13:06 EWI 438 18
11/07/01 12:00 Manual 506 0% 52 10%
11/07/01 12:01 EWI 504 47
08/21/02 13:00 Manual 494 12% 45 86%
08/21/02 13:06 EWI 438 18
DOWNSTREAM
08/17/99 16:30 Blank 22 NA <5 NA
04/19/01 09:00 Blank <50 NA <2 NA
07/09/02 09:00 Blank <50 NA <2 NA
12/12/01 19:40 Replicate 554 3% 110 8%
12/12/01 19:41 EWI 536 102
08/21/02 12:55 Replicate 454 0% 17 11%
08/21/02 12:56 EWI 452 19
03/06/00 09:25 Manual 482 0% 23 4%
03/06/00 09:20 EWI 482 22
11/07/01 11:50 Manual 512 1% 52 2%
11/07/01 11:51 EWI 508 51
08/21/02 12:48 Manual 462 2% 32 51%
08/21/02 12:56 EWI 452 19