Planning evacuations using dynamic fire vulnerability mapping

24 June 2019

USA – In the last three years examples of wildfires in North America that have caused massive evacuations, fatalities, and structures destroyed include:

2016: Fort McMurray, Canada; 88,000 evacuated, 2 fatalities, 3,600 structures.
2017: North Bay, California; 90,000 evacuated, 44 fatalities, 8,900 structures.
2018: Paradise, California; 27,000+ evacuated, 85 fatalities, 19,000 structures

City officials in Paradise had an evacuation plan in place and had even conducted a drill, but the plan assumed a specific fire situation that would allow time for sections of the city to evacuate, one area at at time. The Camp Fire, driven by strong winds, hit the community so quickly that the entire city had to evacuate immediately, causing the limited and low volume evacuation routes to become clogged. A situation like that with very little advance notice would overwhelm many cities, especially if the availability and capacity of routes can’t come close to handling the traffic.

Managers can use computer models to predict the spread of fires, and there are also models that can estimate how much time it would take to evacuate people in vehicles or on foot. But these models have not been integrated to determine how changes in fire behavior would affect evacuation capability and plans.

A linked fire behavior and evacuation model could have variable inputs for weather, fuels, and topography as well as an assortment of evacuation alternatives that could inform planners about existing and proposed designs.

An integrated modeling system or simulator for dynamic fire vulnerability mapping does not exist, but researchers have laid out specifications and a framework for building one. Their recommendations are detailed in a paper published in Safety Science titled, “An open physics framework for modelling wildland-urban interface fire evacuations.”

Illustration from the researchers. Click to enlarge.

Below is the abstract from their paper:

“Fire evacuations at wildland-urban interfaces (WUI) pose a serious challenge to the emergency services, and are a global issue affecting thousands of communities around the world. This paper presents a multi-physics framework for the simulation of evacuation in WUI wildfire incidents, including three main modelling layers: wildfire, pedestrians, and traffic. Currently, these layers have been mostly modelled in isolation and there is no comprehensive model which accounts for their integration. The key features needed for system integration are identified, namely: consistent level of refinement of each layer (i.e. spatial and temporal scales) and their application (e.g. evacuation planning or emergency response), and complete data exchange. Timelines of WUI fire events are analysed using an approach similar to building fire engineering (available vs. required safe egress times for WUI fires, i.e. WASET/WRSET). The proposed framework allows for a paradigm shift from current wildfire risk assessment and mapping tools towards dynamic fire vulnerability mapping. This is the assessment of spatial and temporal vulnerabilities based on the wildfire threat evolution along with variables related to the infrastructure, population and network characteristics. This framework allows for the integration of the three main modelling layers affecting WUI fire evacuation and aims at improving the safety of WUI communities by minimising the consequences of wildfire evacuations.”

Authors of the paper: Enrico Ronchi, Steven M.V. Gwynne, Guillermo Rein, Paolo Intini, and Rahul Wadhwani.

24. June 2019/by GFMCadmin