Rainfall spatial and seasonal variability analysis in semi-arid watersheds

Abstract

A simple procedure for the real-time estimation of the average areal rainfall over a 4 km by 4 km grid was developed. The method proposed here was obtained by a straightforward extension of the well-known block Kriging approach. In order to take into account the seasonal influences, the monthly variogram models were built by least square fit for a ten-year period of rainfall data collected at USDA-ARS Walnut Gulch Experimental watershed. To investigate the rainfall spatial variability on a small scale, a small scale experiment was conducted within the watershed. Sensitivity analysis was also applied between the raingage densities and the relative estimated errors. A portable program MATCORS, for correspondence analysis was developed. A series of graphs and various diagnostic results generated by this program are provided to aid in interpreting the results. It was applied to a monthly rainfall data set recorded statewide in Arizona.