Rainfall erosivity represents the potential of rainfall causing soil erosion. This study aimed to develop a software to estimate rainfall erosivity in the state of Minas Gerais based on Artificial Neural Networks (ANNs). The annual value of the rainfall erosivity is given by the sum of the monthly values of the erosivity indexes EI30 or KE > 25. Two methodologies were used to estimate the kinetic energy for each index. Thus, four erosivity values were evaluated for each month, resulting in the development of 48 ANNs. These ANNs were implemented using the software Borland Delphi 7.0. The new software was called NetErosividade MG. The program calculates the monthly and annual values of rainfall erosivity for any location in the state of Minas Gerais in an easy and fast way.
artificial neural networks; soil conservation; universal soil loss equation
