The objective of the current work is to present a methodology for simulation of stochastic spatial distributed rainfall fields at the daily time step. For this purpose, we develop a geo-stochastic rainfall generating process (SRGP) to generate spatially distributed rainfall fields at daily time scale, that respect the spatial correlation structure of historically observed precipitation, while taking into account important factors that influence the development of observed spatial patterns. For each day, a spatially distributed rainfall field is generated from a pre-specified SRGP, selected based on atmospheric synoptic conditions relevant for that day. Each SRGP is simulated by applying the concept of double kriging, as the product of the spatial amount of rainfall and the spatial occurrence of rainfall by sequential simulation (sequential Gaussian simulation and sequential indicator simulation respectively). The SRGP can account for spatial rainfall nonstationarity related to orographic effects, and can be incorporated as part of a downscaling technique in the context of climate change impact studies. A case study for the Upper Guadiana basin (Spain) is presented that shows the ability of the method to reproduce various spatio-temporal characteristics of precipitation.

Peer Reviewed