An integrated approach of artificial intelligence and geoinformation techniques applied to forest fire risk modeling in Gachsaran, Iran

Forest fires are a multidimensional phenomenon that affects many parts of the world, including the Zagros region of Iran. They are often caused by various factors that can have natural-, anthropogenic-, or combined origins. Considering the significant environmental and socio-economic impacts of forest fires, it is essential to take necessary measures to identify the areas that are prone to forest fires and develop plans and policies for crisis management and risk mitigation accordingly. In this study, we applied an integrated geoinformation (remote sensing and GIScience) approach to analyze and map forest fire risk in Gachsaran, Iran, which is highly prone to forest fires. For the forest fire risk mapping (FFRM), we employed a GIS-based multi-criteria decision analysis method in combination with fuzzy and analytical network process (ANP) methods to identify the forest areas with a high fire risk. To distinguish the vulnerable sites, we employed 13 independent variables encompassing geomorphological factors, land surface characteristics, climatological factors, and anthropological factors. To develop initial criteria maps, we determined the criteria weights using the ANP and used the fuzzy technique for standardization. Finally, the forest fire risk map was produced using the multi-layer perceptron artificial neural network. Our results were also validated against the historical forest fire data using the operating characteristics. Our results showed that 18.417% of the province is subject to a very high forest fire risk. These are areas that should be prioritized when designing precautionary and protective measures. Among the criteria examined in this study, the land surface temperature, soil moisture, and distance from historical forest fire sites received the highest scores in the ANP. The results of this study can be used to identify vulnerable areas, take appropriate planning measures to deal with forest fire risk, and make informed decisions regarding the allocation of facilities in high-risk areas.