Hydrologic Urban Indicators

Science Center Objects

Storm water, the rainfall that runs off urban surfaces such as rooftops, pavement, and lawns, can affect streams in a number of ways. As urban development increases, storm water can run quickly into streams, increasing the volume and peak flows and reducing summer flows. Sediment and other contaminants can also be carried into the streams.

The Washington Department of Ecology (WDOE), the state agency responsible for monitoring local compliance with stormwater regulations and permits, asked the USGS to identify streamflow statistics that WDOE could use to monitor urban streams and the effects of stormwater management efforts. Specific streamflow statistics evaluated in the study will help WDOE monitor the effects of development on streamflow in urban areas.

WA450 - Hydrologic indicators of urban development in western Washington streams - Completed FY2002

Problem - Stormwater production from urban areas modifies streamflow patterns and degrades stream ecosystems. The Washington Department of Ecology (WDOE) has begun Phase I permitting of municipal stormwater discharges in western Washington to address the water-quality problems resulting from stormwater discharges. WDOE has requested assistance from the U.S. Geological Survey to identify streamflow statistics that could be incorporated in National Pollution Discharge Elimination System (NPDES) permits for monitoring the hydrologic conditions of urban streams and the effects of stormwater management efforts.

Objectives - The objective of the proposed project is to analyze the inter-annual variability of three streamflow statistics for streams in western Washington State with long-term discharge records. The project will quantify background variation and trends in the statistics over the period of record. WDOE will evaluate the usefulness of the statistics for inclusion in NPDES permits.

Relevance and Benefits - The hydrologic effects of urban development have been clearly articulated at the storm-scale in terms of peak discharge rate, time-to-peak, and recession rates (Leopold, 1968). These storm-scale measures vary from storm-to-storm and stream-to-stream such that they do not provide a consistent basis for characterizing streamflow patterns in a stream or for comparing streamflow patterns over time. The proposed project will evaluate streamflow statistics to assess their variablity over time under stable land use and their sensitivity to land-use changes. All of the statistics to be analyzed are easily calculated from a continuous record of stream discharges.

The evaluation of streamflow statistics for NPDES monitoring supports the U.S. Geological Survey's goal of improving the usefulness of long-term USGS datasets as articulated under the Environment and Natural Resources mission of the USGS Strategic Plan 2000-2005. The proposed project addresses two priority issues for the Federal-State cooperative program: (1) determining the effects of land-use practices on surface-water quality and (2) evaluating the effectiveness of non-point source pollution management practices (Water Resources Division Informational Memorandum No. 2000.13).

Approach - The proposed project is based on a hydrologic analysis initiated by Konrad (2000) to assess the effects of urban development on annual and inter-annual streamflow patterns. The proposed project will evaluate the temporal variability, including trends, of streamflow statistics for streams in western Washington that have been gaged for at least 20 years. The set will include streams with known land-use change during the period of record and streams with stable land use during the period. Annual values of three streamflow statistics will be calculated for each stream: instantaneous peak discharge rate; the exceedence duration of the mean discharge rate; and 7-day low flow. Background variation will be quantified for each statistic in terms of the coefficient of variation or the 95-percent confidence interval over selected 10-year intervals in all streams and for the period of record in streams where current land use is largely suburban or rural. The variation over the period of record in the statistics will be tested for trends using regression or the nonparametric Mann-Kendall test. Step-trends may be evaluated using Student's t-tests or rank-sum tests.