Abstract

Soil surface erosion is a serious environmental problem which caused decrease in the lands agriculture value and increase in sediment and water pollution. In this research, the aim is to evaluate the supervised classification of ETM⁺ satellite images producing surface erosion map. The Roodbar sub-basin; downstream Sefidrood dam, Guilan province, Iran, was considered as studying basin in a mountainous land type. Results showed that the overall accuracy of surface erosion classification is 69.2%, but accuracy differentiates in different intensities of surface erosion. The greatest accuracy is related to very high intensity and the lowest accuracy is for the low intensity of surface erosion. 

Key words: Surface erosion, erosion mapping, supervised classification, ETM⁺images.