Evaluation of NRCS curve number and MODIS time-series proxies for antecedent moisture condition

Abstract

Soil moisture plays a vital role in a watershed's hydrologic response to a precipitation event. Soil moisture condition antecedent to an event in empirical runoff estimation models, such as the 5-day antecedent moisture model for natural resources conservation service (NRCS) curve number (CN) method, is generalised and spatially inexplicit. This study assesses the potential to parameterise a statistical streamflow estimation model utilising next generation weather radar (NEXRAD) precipitation records and time-series biophysical proxies for soil moisture from moderate resolution imaging spectroradiometer (MODIS) on board Terra satellite. This study is conducted on a 1420 km2 rural watershed in the Guadalupe River basin of southcentral Texas, a basin prone to catastrophic flooding from convective precipitation events. A least squares regression model, accounting for 83% of the variance of observed streamflow for calendar year 2004, was developed based on radar precipitation estimates, land surface temperature, and a vegetation index, on an 8-day interval. Estimated 8-day mean streamflows from the remote sensing model represented an improvement over CN modelled streamflows, developed with the standard and a continuum-based 5-day antecedent moisture condition model. Assessment of remotely sensed proxies of soil moisture has potential for complementing traditional CN method and for estimating streamflow in watersheds for which the CN or other empirical methods are limited by data constraints.