Agent-Based Modeling for Forest Fire Simulation in Chiang Mai, Thailand

Forest fires have a significant impact on ecosystems and human life, especially in northern Thailand, where Chiang Mai is severely affected. Traditional methods for detecting forest fires have limitations, so there is a need for advanced simulation models. This research uses Agent-Based Modeling (ABM) to develop a forest fire simulation for Chiang Mai. The methodology involves collecting historical data, performing Multiple Regression Analysis, designing the simulation, and testing it. Data from 1998 to 2021, including temperature, relative humidity, wind speed/direction, burn area, and slope, were collected from various sources. Multiple regression analysis identified wind speed as the most significant factor affecting burn area. The forest fire simulation, designed using ZF Wang's spread model and tested with AnyLogic, showed that wind direction and speed are crucial in fire spread. The simulation accurately predicted high-risk areas, helping in proactive planning and response. This study confirms that wind is a critical factor in forest fire spread, providing a valuable tool for fire districts to improve preparedness and management. Future research should focus on refining the model with localized data and integrating real-time detection to improve its accuracy and applicability.