|
:
|
Gully erosion is considered to be the dominant type of water erosion in large areas of Iran, especially large parts of the central and southern arid and semi-arid regions of the country. Identifying the effective factors, zoning, forecasting and evaluating areas prone to gully erosion are key factors for land management and preparation. Based on this, the aim of this research is to determine the thresholds of effective factors, zoning and predicting the sensitivity to the gully erosion event in six watersheds of Ilam, located in Ilam province. In this research, a distribution map of the occurrence of ditch erosions in the Ilam watershed was prepared in the form of two random educational (70%) and experimental (30%) groups. Also, the maps of factors affecting the occurrence of gully erosion include general curvature indices, range curvature, profile curvature, convergence, curvature classification, vertical distance from the waterway network, watershed drainage level, waterway power, waterway slope length, topographic humidity, land surface texture, height , slope gradient to percentage, slope direction, shadow and light analysis, density map and distance from the drainage network, lithology, soil type, land use, normalized index of vegetation cover difference, precipitation and climate were prepared based on the available scientific resources in the GIS environment. Multicollinearity test was performed in order to determine and eliminate effective factors that have overlapping information. Then, three models of maximum entropy, Dempstershaffer and witness weight were calibrated (zoning) and validated (prediction) based on the ratio of 70% and 30% of the distribution data of the ditch event in the training and testing phase, respectively. Threshold and relative importance of effective factors in gully erosion were determined based on the superior model (maximum entropy), response curve and jackknife test. The results showed that the factors of land use, lithology, precipitation, soil type, altitude, slope gradient and watershed drainage level are effective in the occurrence of gully erosion, respectively. Also, two curvature factors and waterway power index were removed from the modeling due to their alignment with other factors. Based on this research, the models of maximum entropy, Dempstershaffer and Shahid weight were identified as the most efficient and suitable models for zoning and predicting the sensitivity of land to gully erosion. Considering the similar modeling results of this research in different study areas of Ilam province and the country with hydromorphometric, geological, vegetation and climate features, the results of this research can be used for similar watersheds.
|