TL;DR
This paper presents a novel methodology for integrating storm surge and evacuation modeling to better predict hurricane evacuation clearance times. By coupling advanced storm surge simulations with a transportation evacuation model, the study provides crucial insights into the potential impacts of hurricane track uncertainty on evacuation logistics, using Hurricane Irma in Southeast Florida as a case study.
What Problem Does This Paper Address?
Hurricanes pose a significant threat to coastal communities, and effective evacuation planning is critical for saving lives and minimizing damage. However, accurately predicting evacuation needs and timelines is challenging due to several factors:
- Uncertainty in Hurricane Tracks: Forecasts for hurricane paths often fall within a "cone of uncertainty," meaning the exact landfall location can shift significantly. This unpredictability makes it difficult for both authorities and residents to make timely and informed evacuation decisions. Hurricane Irma's unexpected shift to the west coast of Florida after initially threatening the southeast is a prime example of this challenge.
- Limitations of Existing Models: Current storm surge models often don't account for short-period waves or river discharges, which can exacerbate flooding. Furthermore, evacuation zones are typically static and based on long-term probabilistic forecasts, failing to adapt to real-time changes in a hurricane's track and intensity.
- Traffic Congestion and Roadway Inundation: Large-scale evacuations can lead to severe traffic congestion, further delaying departures. Additionally, storm surge and associated flooding can render critical evacuation routes impassable, creating a dangerous situation for those attempting to flee.
This paper aims to bridge this gap by developing an integrated modeling approach that accounts for both storm surge inundation and evacuation dynamics, specifically addressing the impact of potential hurricane track variations.
Methods
The study employs a two-stage modeling approach:
- Storm Surge and Inundation Modeling:
- A coupled storm surge and wave finite element model, ADCIRC+SWAN, was used to simulate inundation zones.
- This model was forced with Hurricane Irma's hypothetical wind field, assuming landfall in the Miami area.
- The output from this large-scale model was then used as boundary conditions for a smaller, nested inundation model to capture flood water behavior over finer land features.
- The resulting inundation zones and water levels were exported to ArcGIS to identify roadways at higher risk of flooding.
- Evacuation Modeling:
- The inundation data and modified roadway network (accounting for flooded routes) were fed into the Transportation Interface for Modeling Evacuations (TIME) software.
- TIME, which integrates with CUBE Voyager (a macroscopic travel demand model) and CUBE Avenue (for dynamic traffic assignment), was used to estimate evacuation clearance times and shelter availability.
- The model incorporated various scenario assumptions, including population projections for 2020, evacuation levels based on a Category 4 hurricane, shadow evacuations (people evacuating from areas not under mandatory orders), and regional destination rates for evacuees.
This integrated approach allows for a more dynamic and realistic assessment of evacuation challenges compared to standalone models.
Datasets
The study utilized several datasets:
- Demographic and Socioeconomic Data: Obtained from the Statewide Regional Evacuation Study Program (SRESP) and South Florida regional planning councils (SFRPC).
- Shelter Data: Sourced from SFRPC.
- Behavioral Assumptions: Derived from SRESP and SFRPC.
- Hurricane Irma's Track and Intensity Data: Acquired from the National Hurricane Center (NHC).
- Evacuation Zones Data: From NHC.
- Roadway Network Data: From the FDOT Statewide Model Network.
- Population Projections: For the year 2020.
No new benchmarks were introduced; rather, existing datasets were integrated into the novel modeling framework.
Key Insights and Ideas
- The Criticality of Dynamic Modeling: The study highlights that static evacuation zones and models are insufficient. Integrating real-time storm surge predictions that adapt to hurricane track changes is crucial for accurate evacuation planning.
- The "Cone of Uncertainty" is More Than Just a Line: The paper demonstrates how even a slight shift in a hurricane's predicted path can drastically alter the areas at risk and the required evacuation response, as seen with Hurricane Irma's westward shift.
- Roadway Inundation is a Major Bottleneck: Identifying and accounting for flooded roadways is paramount. The study shows that even elevated interstates can be impacted indirectly by surrounding inundation, and minor arterial roads are particularly vulnerable.
- Evacuation Demand is Substantial and Complex: The sheer volume of people and vehicles needing to evacuate, coupled with factors like shadow evacuations and tourist populations, creates immense pressure on transportation networks.
Key Results
- The storm surge modeling indicated maximum water levels on the coast of up to 6 meters, enhanced by waves.
- The integrated model estimated that it would take approximately three days (68 hours) to safely evacuate the populations in Broward, Miami-Dade, and Palm Beach counties under a Category 4 hurricane scenario hitting Southeast Florida.
- Miami-Dade County showed the highest evacuation demand, with over 2 million people and more than 746,000 vehicles requiring evacuation.
- The regional clearance time of 68 hours is remarkably close to the estimated 72 hours for the actual Hurricane Irma evacuation in South Florida, validating the model's predictive capabilities.
Limitations and Future Work
The authors acknowledge several limitations and suggest future research directions:
- Focus on Vulnerable Populations: The current study considered the total evacuating population. Future work could specifically model the evacuation behavior and needs of vulnerable groups, such as individuals with disabilities or the elderly.
- Vehicle Types and Public Transit: The model primarily used total evacuating vehicles. Investigating the impact of different vehicle types, including public transit, on evacuation dynamics would be beneficial.
- Population Projections: While 2020 population projections were used, utilizing the more precise 2020 US Census data could yield more accurate results.
- TIME Model Limitations: The TIME model currently only allows evacuation to designated locations. Expanding this to allow evacuation to other locations would be a valuable enhancement.
- Spatial Accessibility to Shelters: Future research could assess the spatial accessibility of evacuees to shelters, considering shelter capacity and roadway availability.
- Broader Case Studies: Extending the integrated methodology to other geographic regions would further validate its applicability and raise awareness about evacuation decision-making.
- Role of Training: Investigating how training and education can improve evacuation planning and risk reduction is also suggested.
Summary updated