Effects of Raingauge Distribution on Estimation Accuracy of Areal Rainfall

Abstract

Rainfall analysis is important to managing water resources. Mean rainfall is usually used to calculate the spatial rainfall status of a region and is the input into various rainfall-runoff models. However, this method relies on an adequate raingauge network. This study identifies the effects of raingauge distribution based on estimation results of areal rainfall using the Thiessen polygon and block Kriging methods. Twelve rainfall events with complete data from 14 raingauges were selected to complete the goal of this study. The block Kriging method in this study uses a dimensionless semivariogram to obtain hourly semivariograms based on a standardized rainfall depth. The power semivariogram model was used to describe the temporal-spatial variation of rainfall. The analytical process in this study uses raingauge weight and rainfall volume as evaluation criteria. All raingauges were in turn removed from the original raingauge network. The effects of removing each raingauge were compared with computations using all raingauges. Comparison results indicate that (1) the block Kriging method can accurately describe rainfall processes in terms of the spatiotemporal structure of a semivariogram. (2) the block Kriging method is better than the Thiessen polygon method at obtaining exact mean rainfall, and (3) the effects of different raingauge distributions on a mean hyetograph warrant further investigation.