Spatial and Temporal Clustering Analysis of Hotspot Pattern Distribution of Critical Land in Kalimantan, Indonesia

Abstract

Abstract Geographic information systems (GIS) can play a vital function in fire prevention by predicting potential hotspot for fires. This paper presents the analysis of hotspot data in Kalimantan from 2008 to 2018 based on spatial and temporal aspects. Spatiotemporal cluster analysis with Kulldorff’s Scan Statistic (KSS) methods. KSS method is used to explore the spatial temporal patterns. This approach was originally designed to detect clusters and to determine their significance by Monte Carlo replication. the result shows that the provinces with the highest hotspot occurrence cluster are Central Kalimantan and West Kalimantan provinces. Based on critical land area, the cluster distributions of hotspot are dominated in ‘Rather Critical’ and ‘Potential to Critical’ area. Based on land used, the cluster distributions of hotspot are dominated ‘Swamp Shrub’ and ‘Mixed Dry Agriculture Land’. For temporal aspect, the result shows most hotspot occurred in August, September and October. In conclusion, the hotspot detected varied according to year, months and regions. Understanding hotspot patterns enables the allocation of resources for forest fire planning and management. Results showed that the approach is effective for detecting hotspot clusters and cluster accuracy is 91%.