Satellite Rain Estimation in the U.S. High Plains

Abstract

A satellite rain-estimation technique, derived in Florida for convective rainfall, was used to estimate areal rainfall in the U.S. High Plains. Raingages in dense and sparse networks provided the verification data. Unadjusted satellite-inferred rainfalls exceeded the corresponding gage estimates by a factor of 3?5, depending on the area size. This was expected and it is the result of treating convective clouds in arid regions as tropical clouds. Two objective methods were derived to adjust the technique for use in the High Plains. The first involved gage and satellite comparisons for a small area and then extrapolation of this comparison to satellite rain estimates for large areas. The second involved calculation of an adjustment factor using the output of a one-dimensional cumulus cloud model. Accuracy of the adjusted rainfalls are discussed in terms of bias, mean error factor, root mean square error and linear regression analyses. These preliminary results suggest that the satellite convective rain estimation technique can provide rain estimates of considerable utility once the estimates are adjusted for regional differences.