Identification of Eroded and Erosion Risk Areas Using Remote Sensing and GIS in the Quebrada Seca watershed

Abstract

The aim of this research was to identify eroded areas and areas at risk of erosion (EAER) as indicators of soil degradation by water erosion in a semiarid watershed of the Venezuelan Andes in 2017. To this effect, remote sensing techniques and geographic information systems (GIS) were used, focusing on spectral reflectance data from a satellite image, given the absence of continuous pluviographic information and data on soil properties in developing countries. This methodology involved estimating the potential water erosion risk (PWER) and mapping eroded and erosion risk areas (EAER) based on calculating the spectral Euclidean distance to bare soils and a remote sensing technique, which was selected via linear regression. Receiver operating characteristics (ROC) curves were determined to define classification thresholds, which were validated by means of a supervised classification and associated to PWER values. The main results indicate that EAER1 identified more eroded areas with bare soils (229,77 ha) as opposed to EAER2 (195,57 ha). Similarly, it was evident that the first alternative was more successful that the second (sum of the first three principal components). The PWER analysis, in addition to the erosion mapping developed and other data and criteria, such as mini-mum area size of interest, could help to consider necessary soil conservation measures.

Keywords

• spectral euclidean distance
• vegetation indices
• principal components analysis
• maximum likelihood