Identifying areas prone to desertification in Semnan province using logistic regression model (LRM)

Water and soil resources and especially the vegetation cover in the belt areas of the deserts of Iran have provided conditions that cause and intensify desertification and their expansion on the edge of the deserts. Therefore, under the influence of social and economic changes of at least the last century, this sector has progressed towards a critical situation and its size is increasing day by day. Iran is located in the dry belt of the world and a quarter of Iran's area is covered by desert. According to statistics, 18 provinces in Iran are involved in desertification, and one of these provinces is Semnan province. At present, more than half of the province's area is made up of desert lands, which is approximately 53% of the province's surface, the effects of which can be the destruction of water resources, the destruction of groundwater levels, the destruction of wildlife, the increase of respiratory diseases and named .Most of the lands of Semnan province have a dry and semi-arid climate due to their distance from moisture sources, and about half of the lands of the province lack vegetation. Due to excessive exploitation of biological resources such as pastures, forests and water reserves Underground, as well as the illegal transformation of pasture and forest lands into agricultural lands, most of these areas, even the northern lands of Semnan province, have been severely affected by the phenomenon of desertification. In this research, the logistic regression model(LRM) method was used to prepare a map of the vulnerability of Semnan province to desertification. The current method includes the analysis of desert areas with the help of visual observations, measurement, examination and processing of the main indicators using mathematical and statistical models. For this purpose, four independent variables and one dependent variable were used. The independent variables are: vegetation index (NDNI), soil texture index (TGSI) and effective anthropogenic pressure (APSI). and aridity coefficient (AL). Vegetation and soil texture index was obtained from 10 Landsat 8 images. Aridity coefficient was calculated with statistical data of annual rainfall in millimeters and annual evaporation and transpiration in millimeters. The selection of effective human factors in the region has been calculated with the help of AHP method and its map has been prepared. The dependent variable was the Boolean map of areas involved in desertification in Semnan province. With the help of Arc map 10.8.2 and Terrset 2020 software, the coefficient of each of the independent variables has been calculated with the logistic regression method (LRM) and the final risk map of desertification in Semnan province has been obtained. The results of the Regional Soil Texture Survey (TGSI) show that almost more than half of the 60% region has land with high sand content. Among the variables, soil texture factor (TGSI) has been more related to desertification. The results (NDVI) show that almost all the vegetation of the studied area was sensitive to desertification, which was covered by the desert steppe. The desert steppe occupies 99% of the area. The aridity index (Al) is dry in the entire region, except for the northern regions, which include 99.6% of the province. The effect of human pressure in the entire province decreases from the north to the south of the province. The highest pressure of human factors is in the northern region of the province, which covers 41.6% of the area of the province. In order to check the R2 rate, the rate was more than 0.2, which indicates the acceptability of the results of this model. After the classification of the desertification risk map of Semnan Province, which is based on the model (LRM), it was determined that more than 99% of the province's area is at a very severe risk of desertification. The risk of desertification decreases from the south of the province to the north. The north of the province is located in a mountainous region due to its proximity to the Alborz mountain range and the moderate Caspian climate of northern Iran, which has a different climate and vegetation compared to the south. The risk of desertification in the region increases with the decrease in soil quality and vegetation density and weather conditions. The results of the research show the correctness of the LRM model in order to identify areas with a high risk of desertification in Semnan province.