Evaluation of the impacts of land-use change on watershed runoff simulation using the MLP model, Case study: Yalfan catchment

Land-use change affects a number natural processes such as soil erosion, sedimentation, flooding and destruction of soil physical and chemical properties. This change of ecosystem causes the degradation of soil quality which eventually leads to a severe decrease in soil fertility. Therefore, various aspects of land-use change should be taken into account in national major decision makings. This study makes use of Artificial Neural Network, Multilayer Perception (ANN-MLP) and land-use planning model (Dr. Makhdoom’s method) to assess the impacts of land-use change on runoff. The study site was Yalfan catchment, Hamedan province, Iran. According to the results over the statistical period of 1375 to 1389, the Artificial Neural Network model’s (ANN) simulation resulted in the Nash-Sutcliffe model efficiency of 0.74 and root-mean-square error index of 0.01 which was indicative of an acceptable accuracy for the simulation of the daily runoff of Yalfan catchment. Furthermore, the hydrograph peak flow was decreased by 45% after the optimum land-use was implemented. This means a decrease of possible flood damages. It is concluded that the ANN model can be introduced as an appropriate model for runoff simulation in Yalfan and assessing the impacts of land-use change scenarios on runoff.