Valuation methodology of laminar erosion potential using fuzzy inference systems in a Brazilian savanna

Abstract

This study presents an approach on the evaluation of potential laminar erosion in the Ribeirão Sucuri Grande watershed. It is located in the northeast of the state of Goiás, Brazil, a conservation area under strong anthropogenic pressure. A Mamdani fuzzy inference system was designed using linguistic variables, pertinence functions, and a set of rules associated to a traditional laminar erosion prediction model through the environmental conditioners slope, erodibility, and degree of soil protection. The laminar erosion prediction model associated with fuzzy logic is a qualitative evaluation of erosive potential capable of being spatialized with a greater level of detail, increasing the traditional classification by two levels. The processing of environmental and soil conditioning factors using the fuzzy logic resulted in values between 2.5 and 9.1, which places the basin at a low to very high laminar erosion potential. The results indicate areas that demand a greater attention regarding soil management; 56.89% of the area has a medium to high laminar erosion and high to very high erosion (6.99%).