A study on the use and modeling of geographical information system for combating forest crimes: an assessment of crimes in the eastern Mediterranean forests.

This study focuses on the geo-statistical assessment of spatial estimation models in forest crimes. Used widely in the assessment of crime and crime-dependent variables, geographic information system (GIS) helps the detection of forest crimes in rural regions. In this study, forest crimes (forest encroachment, illegal use, illegal timber logging, etc.) are assessed holistically and modeling was performed with ten different independent variables in GIS environment. The research areas are three Forest Enterprise Chiefs (Baskonus, Cinarpinar, and Hartlap) affiliated to Kahramanmaras Forest Regional Directorate in Kahramanmaras. An estimation model was designed using ordinary least squares (OLS) and geographically weighted regression (GWR) methods, which are often used in spatial association. Three different models were proposed in order to increase the accuracy of the estimation model. The use of variables with a variance inflation factor (VIF) value of lower than 7.5 in Model I and lower than 4 in Model II and dependent variables with significant robust probability values in Model III are associated with forest crimes. Afterwards, the model with the lowest corrected Akaike Information Criterion (AICc), and the highest R2 value was selected as the comparison criterion. Consequently, Model III proved to be more accurate compared to other models. For Model III, while AICc was 328,491 and R2was 0.634 for OLS-3 model, AICc was 318,489 and R2 was 0.741 for GWR-3 model. In this respect, the uses of GIS for combating forest crimes provide different scenarios and tangible information that will help take political and strategic measures.