Habitat distribution modeling species ranges southern of Golestan province with Logistic regression

The purpose of this study is to model the distribution of grassland habitats province of Chaharbagh. To this end, sampling of vegetation was systematic-random method that at each site three transects of 50 m was considered. Along each transect, 10 plots of one square meter was placed at a distance of 5 meters.30 plots in each vegetation type and total of 150 plots were carried out. The soil samples were taken at a depth of 0-30 cm. Soil characteristics include the amount of sand, clay content, silt content, lime, gypsum, Ec, pH, electrical conductivity was measured and physiographic factors of slope, aspect and elevation were recorded. After collecting data using logistic regression, the relationship between the plant types and environmental factors were assessed. For predictive mapping of vegetation is necessary to prepare maps of all the factors included in the models, so we used geostatistical methods for mapping soil properties. By the design of the models, with the help of GIS, the predicted distribution maps for each species were obtained. To evaluate the fit of the model predictions with actual map vegetation types, was used kappa index. The coefficient of Artemisia aucheri type equale ./80, The type Festuca.ovina-Astragalus.gossypinus equale ./98, Bromus.tomentellus type equale ./61 And the type Bromus.tomentellus-Festuca.ovina equale ./57, this model showed The type Festuca.ovina-Astragalus.gossypinus had the most accuracy and Bromus.tomentellus-Festuca.ovina type had the least accuracy in the production of habitat maps.