A novel method to identify likely causes of wildfire

Abstract

Natural phenomena, such as wildfires, usually require the coincidence of several related factors in both time and space. In wildfire studies, literature-based factors were collected and listed in Mhawej et al. (2015). The question remains: which combination of factors leads to wildfires? In this context, a novel combination of wildfire likelihood factors was proposed in three different Lebanese forest covers (i.e., pine, oak, and mixed) and related literature-based factors to historical wildfire occurrences. The threshold values of each factor were deduced from the relationship between the element and number of fire occurrences. Each combination of factors was given a unique number. These mixtures corresponded to two, three, four or five factor groupings. The result was the association of each likelihood probability (i.e., low, medium, high, and very high) with different combinations of factors. Ultimately, using these combinations, the wildfire likelihood in Lebanese forests was efficiently and instantaneously generated. This approach could be portable to other Mediterranean regions and applied to several natural hazards.

Keywords

• Wildfire
• Likelihood
• Combination of factors
• Python
• Lebanon
• Natural hazard