Analysis and estimation of urban runoff constituent concentrations as functions of land use characteristics of watersheds at Emporia, Kansas.

Abstract

Eleven watersheds that drain the city of Emporia were investigated to determine relationships between land use and urban runoff water quality. Discharge, conductivity, and pH of dry-weather flow were monitored in each watershed from February through July of 1997. Samples from June and July were analyzed for major nutrients and composited for analysis of major ions and selected metals. The eleven watersheds were mapped and divided into inner, middle, and outer zones based upon distance from the main drainage channels. Land uses and soils in each watershed were mapped. Five data sets for land use and soil coverage were generated by applying five sets of distance-weighting factors to the watershed zones. Each set of land use and soil data was correlated with water quality variables and the five resulting sets of correlation coefficients were compared. The strongest correlations were generally obtained with either the land use data set that gave equal weight to all areas of each watershed or the land use data set that gave weight to the inner zones only. Correlations with concentrations were generally stronger than correlations with rates of loading, and correlations with soil classes were generally weaker than correlations with land use variables. Major ion concentrations had few strong correlations with land use variables but were found to be higher in urbanized areas. pH was lower in urbanized areas, and nitrate concentrations were clearly higher in urbanized areas. Phosphate, potassium, and iron concentrations were linked to vegetated and industrial land uses, and zinc concentrations correlated strongly with railroad land uses. Models of water quality in dry-weather runoff were formulated by regression analysis of constituent concentrations as functions of eight selected watershed variables. Models from the U.S. Geological Survey were used to estimate mean concentrations and total loads of several pollutants in storm runoff for watersheds which drain into the Neosho River and the Cottonwood River, and the estimates were used to model pollutant concentrations in the receiving rivers. Concentrations of nitrate, phosphate, ammonia, copper, and zinc exceeded statutory or suggested water-quality standards in some samples or estimates. A water quality problem at one sampling site was recognized by the consistent absence of macroscopic life.