A GIS based spatial prediction model for human – elephant conflicts (HEC)

Abstract

A spatially explicit model was used to predict the probable human-elephant conflict (HEC) zones in the Agali range of Mannarghat forest division, Western Ghats, India. In the study, various geo-environmental factors such as DEM, slope, aspect, canopy density, distance to forest, distance to hamlets, distance to drainage and distance to river were used for the analysis. Weights-of-Evidence analysis has been applied to find the HEC probability zone using Arc SDM, an extension of Arc GIS. Weight tables were created for each factor, which gave the information about the contrast, showing its positive or negative influence. Using these weight tables, a final predicted map, the posterior probability map was created which predicts the future human- elephant conflict zones. The class (-0.29-0.30) in the ‘canopy density’ shows a contrast value of 2.8515 which is the highest positive influence , while class 1000-1500 (m) in the ‘distance to forest’ shows the least positive influence value of 0.8462. The final posterior probability map shows positive trend towards north-east and negative trends towards south. The relationships between the different zones in the map were compared with, each classes of the eight multiclass geo-environmental variables and land use. We found a strong co-relation of HEC occurrence with distance to hamlets, agricultural land etc.