A Stochastic Model of Radar Measurement Errors in Rainfall Accumulations at Catchment Scale

Abstract

The accuracy of rainfall data is critical for hydrology, as errors in rainfall estimation will affect the quality of hydrological predictions. In hydrology, it is the accuracy of the rainfall measurement at spatial and temporal scales that drive catchment response that is important. Random sampling errors are the residual errors that remain in radar rainfall measurements after systematic biases have been removed. This paper quantifies radar rainfall sampling errors at catchment scale by developing and applying a stochastic space?time model. Random errors caused by temporal and spatial sampling, random variations in the vertical profile of reflectivity, and variations in the reflectivity?rainfall intensity (Z?R) relationship have been included in the model. The stochastic radar rainfall error model is applied in a numerical experiment to quantify sampling errors in radar rainfall accumulations for periods between 10 min and 1 day across catchment areas of between 1 and 1024 km2. Spatial and temporal averaging results in significant reductions in sampling errors at catchment scale. The dominant influence on error in the radar rainfall sampling error model is the height of the radar beam over the catchment.