Analyzing wildfire evacuation dynamics with agent-based modeling in damaged road networks

• Simulates wildfire evacuations with road damage and empirical human response models. • Integrates fire spread, road vulnerability, and traffic flow during evacuations. • Probabilistically assesses congestion points, enhancing wildfire evacuation planning. Wildfires increasingly threaten residents in the Western United States. Despite numerous state and local initiatives aimed at mitigating these risks, completely eliminating the wildfire dangers remains unfeasible due to substantial inherent uncertainties. In this case, evacuation is the most important and effective strategy for reducing human casualties during wildfire events. While the primary goal of evacuation–moving people at risk to safer places–appears straightforward to achieve, the reality is complicated by unpredictable human behaviors and the surge in travel demand, which often results in severe traffic congestion and, consequently, a heightened risk to human lives. In addition, the reduced traffic-carrying capacities of road segments due to wildfires further exacerbate these challenges. In this context, wildfire evacuation simulation can serve as an effective experimental means for emergency management and evacuation planning, offering a cost-effective method to identify bottlenecks and critical congestion points during an evacuation. This paper proposes an agent-based modeling (ABM) framework specifically designed to simulate wildfire evacuations in damaged transportation settings. The proposed framework uniquely integrates wildfire simulation and road network vulnerability assessment with ABM, allowing for a detailed representation of human behaviors during evacuations and the dynamic network functionality in microscopic traffic simulation. A notable contribution of this study is its fully probabilistic approach, which evaluates evacuation performance and identifies critical components of the road network not under a single scenario but under a range of representative scenarios. This probabilistic perspective provides a more comprehensive understanding of potentially vulnerable and congested points, thereby enabling emergency managers and transportation planners to better allocate resources and enhance mobility during wildfire evacuations. The effectiveness of the ABM framework is demonstrated through its application in simulating wildfire evacuations in the City of Santa Clarita, California. The simulation results aid in both pre-fire planning and emergency decision-making, ultimately contributing to improved evacuation strategies and public safety during wildfire events.