This study explored the relationship between increased proportions of imperviousness in a watershed on surface water quality and examined the effectiveness of using remote sensing to systematically and accurately determine impervious surfaces. A supervised maximum likelihood algorithm was used to classify the 2008 high resolution National Agriculture Imagery Program (NAIP) imagery into six classifications. A stratified random sampling scheme was conducted to complete an accuracy assessment of the classification. The overall accuracy was 85%, and the kappa coefficient was 0.80. Additionally, field sampling and chemical analysis techniques were used to examine the relationship between impervious surfaces and water quality in a rainfall simulation parking lot study. Results indicated that day since last rain event had the most significant effect on surface water quality. Furthermore, concrete produced higher dissolved organic carbon (DOC), dissolved organic nitrogen (DON), potassium and calcium in runoff concentrations than did asphalt. Finally, a pollutant loading application model was used to estimate pollutant loadings for three watersheds using two scenarios. Results indicated that national data may overestimate annual pollutant loads by approximately 700%. This study employed original techniques and methodology to combine the extraction of impervious surfaces, utilization of local rainfall runoff data and hydrological modeling to increase planners' and scientists' awareness of using local data and remote sensing data to employ predictive hydrological modeling.
