Influence of Soil Texture Type and Initial Soil Moisture on the Simulation of Seasonal Precipitation and Extreme Precipitation in Southeast Brazil

Abstract

The objective of this paper is to investigate the influence of soil conditions (initial moisture and soil texture type) on the seasonal simulations of precipitation extremes over Upper São Francisco River basin in Brazil. The Eta regional climate model is used to produce simulations for the dry and rainy seasons in southeast Brazil between the years 1979 and 2012. These are 4.5-month integrations. Two maps of soil texture type and two initial soil moisture conditions are tested. Eta rainfall simulations are more sensitive to soil texture than to the changes in the initial soil moisture. The runs using the map of 26 soil texture types represent better the climate of South America compared to the simulations using the map of 9 soil texture types. In general, the extreme climatic indices and the total accumulated precipitation for the dry and rainy seasons in Brazil are underestimated by the Eta regional climate model over the study region, which suggests the necessity of tests with convection precipitation schemes to improve the model’s performance.