Evaluation of nonpoint-source contamination, Wisconsin: water year 1999

Abstract

The objective of the watershed-management evaluation monitoring program in Wisconsin is to evaluate the effectiveness of best-management practices (BMPs) for controlling nonpoint-source pollution in rural and urban watersheds. This progress report provides a summary of the data collected by the U.S Geological Survey for the program and a discussion of the results from several different detailed analyses conducted within this program. A land-use and best-management-practice inventory is ongoing for each evaluation monitoring project to track the different sources of nonpointsource pollution in each watershed and to document implementation of best-management programs that may cause changes in the water quality of streams. Updated information is gathered each year, mapped, and stored in a geographic-information-system database. Summaries of BMP-implementation data collected through the 1999 water year are presented in this report. Suspended sediment and total phosphorus storm-load and annual-load data are summarized for eight rural sites. For all 8 rural sites a sufficient number of pre-BMP storm samples have been collected; for two of the sites (Brewery and Garfoot Creeks), a sufficient number of post-BMP storm samples have been collected to allow for a final assessment of the effectiveness of the BMPs. For the remaining sites, numerous transitional storm samples have been collected, but in all cases BMP implementation has lagged such that there are insufficient post-BMP storm samples for final analysis. For two sites (Rattlesnake and Kuenster Creeks) there are not enough planned BMPs to warrant further data collection. Continuous dissolved-oxygen data collected at 5 rural sites are summarized. In terms of instantaneous concentrations when comparing pre-BMP data to transitional and post-BMP data, the general trend is a reduction in the number of days that the dissolved oxygen concentration was less than the state standard. These results are anecdotal, however; the differences have not been rigorously tested statistically. For a level of dissolved oxygen sustained over a continuous hour, the results are mixed. In general the number of days with standard violations has decreased, but there are notable exceptions. For the four urban streams, the pre-BMP data were examined to determine the level of improvement that could potentially be detected with a statistical analysis. Regression analyses were performed relating constituent loads of suspended solids, total phosphorus and total copper to various independent variables, including seasonal terms and variables related to rainfall. On the basis of the residuals from the regressions, there is a wide range of potential change that could be detected with an analysis of pre- and post-BMP loads. This is likely a result of the high degree of variability in the data, particularly from site to site. For suspended solids, total phosphorus, and total recoverable copper the minimum detectable changes ranges from 20-80, 30-70 and 30-90 percent, respectively. For two of the eight rural streams (Rattlesnake and Kuenster Creeks) minimal BMP implementation has occurred, hence a comparison of pre- BMP and data collected after BMP implementation began is not warranted. For two other rural streams (Brewery and Garfoot Creeks), BMP implementation is complete. For the four remaining rural streams (Bower, Otter, Eagle, and Joos Valley Creeks), the pre-BMP load data were compared to the transitional data to determine if significant reductions in the loads have occurred as a result of the BMP implementation to date. For all sites, the actual constituent loads for suspended solids and total phosphorus exhibit no statistically significant reductions after BMP installation. Multiple regressions were used to remove some of the natural variability in the data. Based on the residual analysis, for Otter Creek, there is a significant difference in the suspended-solids regression residuals between the pre-BMP and transitional periods, indicating a potential reduction as a result of the BMP implementation after accounting for natural variability. For Joos Valley Creek, the residuals for suspended solids and total phosphorus both show a significant reduction after accounting for natural variability. It is possible that the other sites will also show statistically significant reductions in suspended solids and total phosphorus if additional BMPs are implemented.