Modeling soil erosion processes in watersheds and the relation between soil loss with geomorphic and chemical parameters

Soil erosion by water has led to severe economic and environmental impacts and remains a global issue because of its extent and complex processes. Loss of soil nutrients, declining crop yields, and reduction in soil fertility are only a few problems that caused by soil erosion. In watershed areas soil, moved by erosion, carries dangerous chemicals into rivers, streams, and groundwater resources, while simultaneously it causes a reduction in water-bearing capacity of an aquifer as well as in the groundwater quality. An effort to understand soil erosion processes is carried out in the present study. In a first step, the Revised Universal Soil Loss equation (RUSLE) model is adopted in a Geographical Information System (GIS) framework in order to predict the potential annual soil loss for several major watersheds of Crete Island, southern Greece, where soil erosion is a growing problem. The RUSLE factors were calculated (in the form of raster layers and in a cell by cell basis) for seven major watersheds which cover the northeastern and central part of the Chania and Rethymnon Prefectures, respectively. The R-factor was calculated from monthly and annual precipitation data. The K-factor was estimated using the soil data available from the Soil Geographical Data Base of Europe at a scale of 1:1,000,000. The LS-factor was extracted from a 20-meters digital elevation model (DEM) of the study area. The C-factor was calculated through the Normalized Difference Vegetation Index (NDVI) approach using Remote Sensing (RS) techniques while the Practice Support factor was set to 1. The results show that an extended part of the area is undergoing severe erosion. The mean annual soil loss is predicted more than ~200 (t/ha yr) for Kalami and Argyroupolis watersheds while Vrysses watershed exhibits annual soil loss more than 300 t/ha yr, showing extended erosion and demanding the attention of local administrators. In a next step, possible relation of soil loss raster layers (with a cell size of 20*20 meters) spatial distribution with structural, geomorphic and chemical parameters of the watersheds is investigated. Specifically, data derived from Remote Sensing (such as NDVI), interpolated point data (e.g., borehole data, chemical data) slope angle, geological and permeability data are fully integrated in a GIS environment. Ordinary Least Squares linear regression is finally applied for analyzing the relationships between the soil loss and the unique parameters. Finally, the obtained results provide information for an improved understanding of soil erosion processes in watersheds.