FLEE: A Natural-Human Model To Explore How Hurricane Forecast Errors Affect Evacuations

Coupled natural-human models offer potential to bridge the physical, social, and computational sciences to support the weather-climate enterprise. One of these models, FLEE (Forecasting Laboratory for Exploring the Evacuation-system), uses an agent-based modeling approach to explore how hurricane forecasts interact with society and the built environment to impact evacuation decision-making and traffic. This presentation highlights key features of FLEE, including recent improvements to the model’s resolution, mapping of real-world data, and representations of hurricane risks, that enhance FLEE’s capacity to connect NWP with societal outcomes. Then FLEE’s simulated evacuations – specifically its evacuation orders, evacuation rates, and traffic – are compared to available observational data on evacuations during Hurricanes Irma, Dorian, and Ian. Lastly, several proof of concept experiments highlight FLEE’s capacity for exploring the relationships between forecast errors and evacuation outcomes. In summary, this work demonstrates FLEE’s capacity as a virtual laboratory and a “societal impact tool” for the meteorological community, taking a step towards making weather forecasts even more valuable to society.