Techniques for estimation of storm-runoff loads, volumes, and selected constituent concentrations in urban watersheds in the United States

Abstract

Introduction 1 Purpose and scope 2 Acknowledgments 2 Data base 2 Estimating procedures for storm-runoff loads and storm-runoff volumes 4 Methods 4 Selection of response and explanatory variables 4 Definition of homogeneous regions 5 Selection of model form 6 Models 7 Three-variable models 7 Comparisons of all storm-runoff-load and storm-runoff-volume models 7 Storm-runoff-load models for region 1 9 Storm-runoff-load models for region II 11 Storm-runoff-load models for region III 12 Storm-runoff-volume models 12 Three-variable models for storm-runoff loads 14 Limitations of significant explanatory variables 17 Other potentially useful explanatory variables 18 Validation, testing, and application of regression models 19 Split-sample analysis 19 Standardized beta coefficients 20 Application of regression models 20 Estimating procedures for storm-runoff mean concentrations 21 Methods 21 Models 21 Estimating procedures for mean seasonal or annual loads 22 Response variable-mean load for a storm 22 Explanatory variables 23 Methods 25 Models 28 Example 31 Summary 31 References cited 32 Contents III Contents VI Manuscript approved for publication, March 17, 1988. application in the more arid Western States, and the least accurate models were those for areas that had large mean annual rainfall .

Verification and adjustment of regional regression models for urban storm-runoff quality using data collected in Little Rock, Arkansas

“…Mean seasonal load also can be estimated by calculating P^ for a single storm representing average seasonal conditions for TRN, then multiplying P^ (the average seasonal storm load) by the average number of storms per season (Driver and Tasker, 1990). The mean annual load also can be estimated by summing the mean seasonal loads.…”

“…Some of the prediction error is attributed to error in the regional models due in part to differences in physiographic settings. In the NURP region III data base, 8 of the 11 cities are in, or very close to, a coastal setting (Driver and Tasker, 1990, figure 1). Furthermore, most of the cities in the NURP region III data base are larger and have been established longer than Little Rock.…”

“…Predicted values of storm-runoff load and mean concentration for each of the monitored storms were computed from the linear regression models for region III presented in table 2 (Driver and Tasker, 1990). Region III consists of areas within the United States that have mean annual rainfalls equal to or greater than 40 inches.…”

“…Rainfall, discharge, and water-quality data were collected during representative storm events between June 1992 and January 1994, at five representative catchment areas in the city of Little Rock. The local data collected were used to verify and adjust regional regression models for estimating storm-runoff loads and concentrations (Driver and Tasker, 1990). A model adjustment procedure (Hoos and Sisolak, 1993) was used to adjust the selected regional models for use in the Little Rock area.…”

Nutrient sources and loads in the Connecticut, Housatonic, and Thames River basins

“…Urban storm runoff loads are affected by total storm rainfall, total contributing drainage area, peak discharge, storm runoff volume and duration, antecedent rainfall conditions, rainfall rates, percentage of impervious area in the drainage basin, land-use characteristics, mean annual climatic characteristics, and other factors (Driver, 1989;Driver and Tasker, 1990). Because of the wide range of factors affecting urban nonpoint runoff, additional information for specific locations and conditions is necessary to adequately characterize the expected range of urban nonpoint contributions to streams, and to quantify the contributions of different urban land uses.…”