To search, Click below search items.


All Published Papers Search Service


Gully erosion Sufficiency mapping at Robatturk Watershed (Iran) using an artificial neural network model


Heshmatolah Agharazi , Ali Aakbar Davoudirad, Saeed Khosrobagi, Samad Shadfar, Shamsolah Nikchah, Abolfazle Najimi


Vol. 17  No. 4  pp. 14-20


Spatial prediction of occurrence of gully erosion through the use of models that are based on the sufficiency of land to gully erosion and output them to the hazard mapping of gully erosion in the gully erosion hazard zone, most appropriate strategy for land management planning in watersheds prevent the occurrence of erosion. Therefore, in this research, the sensitivity of gully erosion in the Robatturk watershed Markazi province of Multilayer Perception neural network structure and the use of variables the selected suitable factors are: lithology, land use, distance from river, distance from road, soil texture, slope degree, slope aspect and altitude. The results of ANN represents the final structure 1- 6-8 with 0.5 of learning and sigmoid activation function in the hidden layer for mapping is appropriate to the sensitivity of gully erosion in this area. The result of erosion hazard zone using artificial neural network using the 1-6-8 structure, learning about 0.5 and sigmoid activation function in the hidden layer shown that 70.35% very low, 3.10% in low-class, 1.05% percent in the medium risk category, 4.14% and 22.39% are classified as high risk is too much risk.


ANN, hazard zone, Landuse, Markazi Province