Urban stormwater runoff pollutant loadings: GIS land use classification vs. sample-based predictions

Abstract

Cities are growing worldwide with an increase in stormwater quantity and decrease in quality, negatively impacting receiving water bodies. The characterization of stormwater is difficult given its high variability and the typically numerous outfalls to be monitored. However, loadings can be estimated via models and validated using actual outfall monitoring. This study determined stormwater pollutant loadings predicted using eight land-use classifications (i.e., a 'desktop' study) and via an outfall sampling regime (i.e., a 'monitoring' study) for seven stormwater catchment areas in Saskatoon, SK, Canada, where stormwater typically releases directly into the South Saskatchewan River. Pollutants considered were total suspended solids (TSS), chemical oxygen demand (COD), metals, and polycyclic aromatic hydrocarbons. Catchment areas were dominated by single-family residential (39%) and green areas (17%). The largest catchment area, Preston Crossing, was the major source of the predicted annual loadings, such as TSS at 550,000 kg and COD at 265,000 kg. For comparison, the sampled-based estimated loadings for TSS and COD were 362,700 kg and 652,700 kg, respectively. Differences between the average predicted and actual estimations ranged from 29 to 156% for the eight pollutants considered, with averages for the summed pollutants in each catchment area ranging from 48 to 130%. Overall, the assessment and monitoring of stormwater outfalls are needed for the determination of impacts of loadings on the environment and for the subsequent development and implementation of treatment technologies.

Keywords: GIS; Land-use classification; Pollutant loads predictions; Polycyclic aromatic hydrocarbons (PAHs); Stormwater quality, pollutant emission; Stormwater runoff.