Examination of aggregation effects in watershed hydrological modeling using microwave-derived soil moisture data

In a previous study, the usefulness of incorporating microwave-derived soil moisture information in a statistical version of a hydrological model was presented. In the present study the TOPLATS hydrological model was initialized with surface soil moisture fields derived from the ESTAR airborne L-band microwave radiometer for the Washita '92 experiment in the Little Washita River watershed in Oklahoma. After only a single initialization with ESTAR data on the wet day at the start of the study period, the hydrological model was able to produce more accurate predictions of soil moisture than a standard hydrological initialization with streamflow data over an eight-day soil moisture drydown. However, when the model soil moisture was compared to ESTAR soil moisture on each of the eight days, the spatial patterns of the model errors were not random-soil moisture was overpredicted in the sandy soils in the center of the watershed while underpredicted in the loamier soils to the east and west. Since the statistical version of the hydrological model is driven by the watershed average of the hydrological parameters, it is difficult to conclude whether these modeling results are due to errors in the input hydrological parameters obtained from a soils data base or due to errors resulting from the aggregation processes in the model. Thus, the objective of this paper is to use the fully distributed version of the hydrological model to explore the detailed representation of the dynamics of hydrological processes in the watershed, so that the effect of aggregation at the macroscale can be determined.