Rainfall seasonality and erosivity for two agro-industrial regions in Brazil

Abstract

The aim of this study was to determine the rainfall erosivity (R), return period and the probability of occurrence of rainfall in Cáceres and Rondonópolis municipalities in the state of Mato Grosso, Brazil. The rainfall erosivity index (EI30) and rainfall coefficient (Rc) were calculated using pluviometric data for the periods of 1968-2009 and 1970-2009 for the stations of Rondonópolis and Cáceres, respectively. Equations were then fit in order to estimate the monthly EI30 values from monthly Rc for each station. The study show that EI30 values estimated from equations relating rainfall coefficient and erosivity for a particular place, even if obtained from a small series of pluviometric data, are more reliable than the values estimated from equations of nearby locations. The proximity between two locations should not be used as a unique condition to extrapolate the values of EI30m and Rcm since this proximity is not necessarily related with the similarity of the physical characteristics of rainfalls. The results show good relationship between EI30 and Rc, providing, for Cáceres and Rondonópolis, R estimates with high reliability. The stations of Cáceres and Rondonópolis have R average values of 5,579 and 6,904 MJ mm ha-1 h-1 year-1, a return period of 2.1 and 3.7 years, and probability of occurrence of 46.8 and 26.9%, respectively. For both stations, the months with most erosive rainfall are december, january and february, which correspond to 45% (Cáceres) and 52% (Rondonópolis) of the annual erosivity.