A SPATIAL MODEL FOR PREDICTING THE OCCURENCES OF DEFORESTATION IN THE ISLAND OF SUMATRA, INDONESIA

Abstract

This study primarily aims to develop a spatial model for predicting the occurrences of deforestation in Sumatra Island, Indonesia. This study was conducted based on spatial logistic regression approach that has been widely acknowledged for its flexibility and ability to accept a mixture of both categorical and numerical variables. Result of this study shows that a combination between logistic regression-based modelling and Geographical Information System (GIS) is indeed suitable for determining the probability of deforestation occurrences in Sumatra Island. Analysis conducted in this study has also revealed that physiographic variables, soil type variables, as well as human activity variables have high significant correlation with deforestation. These findings are useful to assist policy makers in Indonesia to understand the process of deforestation and to take it into consideration while formulating land use-related decisions.