The U.S. Environmental Protection Agency requires information on the volume and quality of urban storm-water runoff to apply for a permit to discharge this water into the Boise River under the National Pollutant Discharge Elimination System Program. Concentrations of selected chemical constituents in storm runoff were determined from samples collected at four storm-sewer outfalls in Boise from October 1993 through June 1994 and at one outfall in Garden City from September through October 1994. Samples were analyzed for specific conductance, pH, alkalinity, water temperature, oxygen demand, fecal indicator bacteria, major ions, dissolved and suspended solids, nutrients, trace elements, and numerous organic compounds. The measurement of storm-runoff volume and mean concentrations of constituents were used to estimate storm-runoff loads. Previously developed U.S. Geological Survey regional regression models of runoff and 11 chemical constituents were evaluated to assess their suitability for use in urban areas in Boise and Garden City. Data collected in the study area were used to develop adjusted regional models of storm-runoff volumes and mean concentrations and loads of chemical oxygen demand, dissolved and suspended solids, total nitrogen and total ammonia plus organic nitrogen as nitrogen, total and dissolved phosphorus, and total recoverable cadmium, copper, lead, and zinc. Explanatory variables used in these models were drainage area, impervious area, land-use information, and precipitation data. Mean annual runoff volume and loads at the five outfalls were estimated from 904 individual storms during 1976 through 1993. Two methods were used to compute individual storm loads. The first method used adjusted regional models of storm loads and the second used adjusted regional models for mean concentration and runoff volume. For large storms, the first method seemed to produce excessively high loads for some constituents and the second method provided more reliable results for all constituents except suspended solids. The first method provided more reliable results for large storms for suspended solids.