Analysis of adjustment of satellite precipitation for streamflow simulations

Abstract

The lack of dense ground networks of meteorological stations in many parts of the developing world impedes accurate hydrological modelling. These are regions that could greatly benefit from satellite precipitation data. However, there are significant problems with satellite products in providing reasonable estimates of precipitation over land. Methods exist that can adjust satellite products (with ground gauge or radar data) providing more robust estimates of precipitation. However, high quality ground data may not exist within the pixel, typically 0.25 o x 0.25 o , to provide the necessary information to facilitate the adjustment of the satellite product. To address this problem this study will examine how streamflow performance, based on satellite products (TRMM 3B42 Real-time -TMPA-RT; CMORPH) adjusted with ground precipitation data, will vary over a range of spatial transfer distances (0 o to 3.00 o ). The spatial transfer distance is the mean distance that a rain gauge is located away from the centroid of the watershed. This study compares simulated (driven by adjusted satellite precipitation) and observed streamflow from six moderately large basins from the United States. These basins span a range of climatic conditions from dry (San Pedro Basin -Arizona; Cimmaron Basin -Oklahoma; Nueces Basin -South Texas; middle Rio Grande Basin -Texas and northern Mexico) to humid (Alapaha Basin -Georgia; Upper Tar Basin -North Carolina). This study is unique because it quantities how far spatially transferred precipitation data can be potentially applied to support hydrologic modelling, which is knowledge that can be applied to poorly gauged regions of the world.

Temporal analysis of Soil and Water Assessment Tool (SWAT) performance based on remotely sensed precipitation products

“…Comparison of precipitation products and streamflow simulations (at daily and monthly time scales) based on different precipitation products for six of seven examined basins have been previously published (Tobin and Bennett, , ,, , ). In brief, the average annual precipitation values for products that support acceptable streamflow simulations are tightly clustered with each other as indicated below.…”