استفاده از روش‌های آماری رگرسیون دو و چند متغیره در پهنه بندی خطر زمین‌ لغزش (مطالعه موردی: حوزه آبخیز شمال تهران)

During recent years, landslide incidence has been increased in different areas with different reasons like change in land use and road introduction. It is necessary to research and survey that’s creation factors and identity sensitive areas to prohibit of probable damages. For this reason, This study were carried out in Northern of Tehran watershed with about 42.14 km2 area. In order to landslide hazard zonation in this area, Two and Multi Variate Regression methods were selected after survey the watershed. At first, nine effective factors on landslide were identified. Including: slope, aspect, height, distance from fault, lithology, distance from stream, distance from road, Precipitation and land use. Then, data layers using Arc GIS base on each methods were prepared. Then the land slide hazard zonation maps were prepared.  For evaluation methods, the land slide hazard zonation maps of each method were compared with actual landslide scattering maps. The results of two variate regression shows that occurrence of landslides have significant relevance with distance from road, slope, distance from fault, precipitation and lithology parameters, that R2 measures are 0.51 for distance from road, 0.31 for slope, 0.167 for distance from fault, 0.33 for precipitation and 0.10 for lithology. R2   measure of multi regression was 0.65 that the nine factors were significant at 0.99 level. The results showed that (P) measure in two and multiregression methods are 0.89 and 0.92, respectively. According to results the multi regression is better than two regression method.