A Multi-Source Reconstruction of the 2024 Jasper Wildfire Complex Evacuation and Reentry Periods

• Jasper (Canada) was evacuated for twenty-four days due to a severe wildfire in 2024. • Events were reconstructed via network mobility data, traffic counts, and social media. • Transportation network characteristics severely constrained evacuation flow. • The evacuation of Jasper took eight hours, and reentry lasted for multiple weeks. • Evacuation departure curves were similar to Poisson, logistic, and Rayleigh functions. Wildfire events pose extreme danger to residents, requiring evacuations to safety. To better inform evacuation planning, a post-disaster reconstruction of events using multiple sources can identify emergent travel behaviour, evacuation time estimates (ETEs), departure timings, and flow characteristics. In this study, the 2024 Jasper Wildfire Complex was reconstructed with de-identified network mobility data from the TELUS Data for Good Program, open-source traffic counts from the Alberta Government, and official government correspondence via social media platforms. By leveraging unique insights from three different data sources, we develop a more complete understanding of the evacuation across the entire duration of the wildfire compared to single-source studies. The results found value in operating a unified command between municipalities and counties, a peak hours-long evacuation flow estimate of 800 vehicles per hour on a one-lane rural highway, and a steady rate of 2,000 device departures per hour that best matched Poisson, logistic, and Rayleigh functions. The evacuation of the Jasper area was estimated to take eight hours, limited by the single-lane egress route, indicating that a combination of transportation strategies is required for timing reductions in similar jurisdictions. A slow, weeks-long rate of return was observed during reentry, indicating the importance of long-term community supports and resources. Positively, strong similarities between network mobility data and traffic counts showcase their data validity in investigating future evacuation events. The results are most applicable to rural communities with a single egress route and a similar hazard context as Jasper.