The effects of cropping systems on soil erosion risks and crop productivity using ImpelERO model and GIS in northeast of Iran

Abstract

Soil erosion is a severe problem worldwide, and controlling it will be a major challenge in the future. Neural networks, as an artificial intelligence technology, have grown rapidly over the past few years and have an ability to deal with nonlinear multivariate systems. An integrated Model to Predict European Land use named ImpelERO is a decision trees/neural network hybrid model. The overall approach of ImpelERO was applied in 49 selected sites from Mashhad–Chenaran plain to quantify the soil erosion features including soil vulnerability index, soil loss rate, erosion risk class and soil depth reduction by sugar beet cultivation under conventional and conservational management practices. Our results revealed that the soil vulnerability indexes ranged from 0.12 to 0.54 and 0.1 to 0.44 by conventional and conservational practices, respectively. The values of soil losses in the study area varied between 4 to 59.7 t ha−1 year and 3.4 to 38.7 t ha−1 year with an average of 11.72 and 7.83 t ha−1 year by conventional and conservation management practices, respectively. The mean values of erosion risk classes ranged from V3 in conventional to V2 in conservational practices which categorize the region as accepted tolerable to sensitive to erosion. The long term soil productivity reduction for time horizons 2020, 2050 and 2100 revealed that the conservational practices have greater contribution on preventing productivity reduction than conventional management practices.