Exploring and forecasting spatial and temporal patterns of fire hazard risk in Nepal's tiger conservation zones

Forest fires are the leading cause of deforestation. In March, April, and May, which are the three driest months of the year, 89% of all forest fires in Nepal are caused by humans. Due to its diverse topography, climate, and sociodemographic interactions, the region is extremely vulnerable. The study aimed to investigate and forecast the spatiotemporal pattern of forest fires using space-time cube forecasting models and develop a tiger habitat suitability model based on space-time hotspots, trends, environmental, geomorphological, and human components. MODIS fire hazards data in vector point format from NASA's Terra and Aqua satellites between the years 2000 and 2020 were utilized as the primary dataset for this investigation. Fire hazards from 2000 to 2013 were utilized to analyze the spatiotemporal hot-cold spots trend, hotspot zone in 2D and 3D. Then fire hazards from 2014 to 2020 were forecasted and validated with raw datasets using the space-time exponential-smoothing and forest-based forecasting models. Results indicate that almost 52.77% and 69.05% of the total tiger habitat area were respectively in hotspot and uptrend zones. Only 0.3% of the entire tiger habitat area was identified in a cold-spot zone. Multi-Criteria Evaluation (MCE) suitability analysis estimated that the areas of Mahakal and Seti in the west and the top of Mechi in the east were found to be moderately suitable for tiger habitat. When compared to the exponential smoothing method (8.40%), the forest-based method produced a lower average inaccuracy of 8.29%, with mean RMSEs of 0.43 and 0.31, respectively. Overall, the study revealed a new era of technological spatiotemporal data utilization for fire hazard incidents, which could be used to identify suitable or vulnerable locations through space-time analysis and forecasting techniques.