Analysis Of Fire Risk For Forest And Land In West Kalimantan Using Logistic Regression Method With Generalized Extreme Value Approach

Abstract

Forests in Indonesia have been reduced by half due to fires. Forest and land fires often occur during the long dry season in places such as the island of Borneo. West Kalimantan is an area passed by the equator which is directly above the Pontianak area. The main effect is to make West Kalimantan a tropical area with high air temperatures so that forest and land fires often occur. This study aims to obtain the results of the probability of land and forest fires in each district in West Kalimantan. The method used is binary logistic regression analysis with response variables in the form of data categories based on spatial data and analysis of extreme values with Generalized Extreme Value (GEV). Spatial analysis uses the help of ARCGIS software in processing raster data (grid cells). The data used is data on maximum temperature and maximum wind speed taken from October 7, 2021 to October 31, 2022 from the official NASA website. The spatial data used in this study is forest and land fire vulnerability data taken from the BNPB website in the form of raster data. The results of logistic regression analysis found that the maximum temperature variable has a negative relationship with the response variable, while the maximum speed of wind variable has a positive relationship with the response variable. The temporal probability of the resulting GEV is getting higher with a longer period of years ahead. The probability of forest and land fires is obtained by multiplying the log probability by the GEV temporal probability. In this study, it was found that the highest chance of forest and land fires occurring in Sanggau Regency was suspected to occur due to an increase in temperature every year.