Use of a Multiple Linear Regression Model to Estimate Stormwater Pollutant Loading.

Abstract
It is widely hypothesized that the quality of stormwater runoff varies by land use. Therefore, pollution loading from a particular watershed can be formulated as a function of the area of the various land use types that form it. In order to develop an estimate of regional pollutant loads, it is desirable to derive a best estimate of the quality of storm runoff from different land uses in the region. A multiple linear regression model was developed from water quality and land use data collected during the 1989-90 wet weather season in Alameda County, California. This model was used to estimate runoff pollutant concentrations by land use for this region. The resultant pollution concentrations were used to estimate the pollutant loads for an average year. Both single storm event data and averaged data from eleven storm events occurring during one wet weather season were used. The F-test was applied to statistically test the validity of the model.
This paper is only available in PDF Format: View full text PDF