Determining relationships between soil properties and plant distribution in a protected area in central Iran

A hybrid algorithm specifically designed to work with optimised support vector machine with genetic algorithm (GA-SVM) was developed for determining the relationships between soil properties and plant distribution and vegetation cover densities in a protected area (Ghomeshlu, central Iran). The bulk density, porosity, silt, total nitrogen and chloride contents are the main essential factors (with a screen accuracy of 100%) for the establishment and growth of Scariola. For Astragalus, surface fragment content has the greatest influence, while available phosphorus was screened by the GA-SVM analysis as the factor with a closer relationship with Anabasis growth in the study sites. Particle density, aggregate stability, available magnesium and pH are the more important combination of soil properties affecting the coverage density of Stipa. Soil organic matter content, available phosphorus, total nitrogen, electrical conductivity, porosity and particle density have a closer relationship with the coverage density of Noaea. This study provides a strong basis for identifying habitat characteristics of vulnerable and/or endangered species in Iran.