Integrating GIS and remote sensing approaches with NEXRAD to investigate the impacts of spatial scaling of hydrologic parameters on storm-runoff modeling

Techniques of Geographic Information Systems and remote sensing are integrated with NEXRAD weather radar data in a watershed modeling package, HYDROGIS, to investigate the spatial scaling of hydrologic parameters and their impacts on the accuracy and spatial distribution of surface runoff. Precipitation data estimated by both rain gages and NEXRAD radar, land use/land cover information derived from SPOT XS image, and the Soil Conservation Service (SCS) Curve Number methods were used to model the Little Washita river basin, a rural watershed in south-central Oklahoma. Storm runoff processes were simulated at various spatial scales using four different aggregation methods (majority, median, average, and image resampling) to estimate the accuracy and spatial reliability of runoff at each aggregating level.; Results indicate that spatial structure of landscape complexity has a significant impact on the spatial scaling and high runoff-generating areas play an important role in surface runoff processing while the critical scale for this particular basin is about 300 m. Spatial patterns of soil and precipitation have less impact on spatial scaling at scales of less than one kilometer although this result occurs, in part, due to the fact that they were measured at a coarser resolution. In addition, the method employed to aggregate data or parameters is of much importance--of the four approaches investigated, the image resampling method was the best one according to both accuracy and spatial reliability while average method is the worst.