Evaluating the benefits of spatio-temporal population dynamics data for protective action decision support in wildfire emergency management

Knowledge of dynamic locational social aspects in terms of varying population distribution and vulnerability patterns is crucial for decision support in proactive disaster risk as well as immediate situational crisis management and thus considered one of the key factors for strategic planning and disaster impact minimization. This is relevant in the pre-event phase for designing appropriate preparedness measures and evaluating required resources and even more so during acute crisis situations for efficient timely and situation-aware response. Situation-specific decision-making in disaster risk management including crisis response is strongly context- and therefore event- and hazard-related. In particular this refers to the basic difference in the speed of onset, the duration and the dynamic development of an event. For example, earthquakes are characterized by very short onset and short durations, whereas heat waves and droughts as well as flood events usually imply a certain lead time in their development and are also slow to fade away. Forest fires usually don´t affect a populated area entirely without warning time. Nonetheless fires can progress rapidly under certain circumstances and threaten areas of human activity with very limited preparation and reaction time. Potential threatening situations can appear suddenly,trigger an entirely new hazard area. Obviously, the development of wildfire emergency situations is therefore often very hard to predict which makes a dynamic management understanding especially for response actions indispensable. Fire emergency managers need to be able to identify the likelihood that wildfire will affect valuable resources in real-time. Focus in that regard lies on guaranteeing population safety first of all as well as protecting private structures, public infrastructure, as well as natural and cultural resources. With regard to the consideration of population distribution and density patterns, while in an earthquake context it is essential to have information whether people are inside or outside of buildings at the time of the shock (collapsing buildings as primary source of related casualties), for forest fire disaster risk management mainly the `mere´ geographic location of potentially exposed people is relevant. That geographic location of population clusters is then one of the basic parameters for protective action decision support. Wildfire related casualties usually happen when people get trapped in an area where exits are blocked by the approaching fire and protection status of the local structures/homes is not sufficient considering the event´s intensity. Also in that case it is crucial to have a clear status on how many people are located in such areas with those affected areas varying rapidly and dynamically during the event when the fire progresses. This paper comes up with a conceptual framework on how spatio-temporal population Dynamics models such as the currently developed DynaPop can benefit pro-active decision making in a wildfire emergency management context. Relevant strategies include evacuation and shelter-in-place whereby temporal aspects are crucial throughout the entire event action chain due to the highly dynamic nature of wildfires and the variable and quickly changing decision-relevant factors regarding community exposure and vulnerability.