Abstract For this study, we present and evaluate an improved agent-based modeling framework, the Forecasting Laboratory for Exploring the Evacuation-system, version 2.0 (FLEE 2.0), designed to investigate relationships between hurricane forecast uncertainty and evacuation outcomes. Presented improvements include doubling its spatial resolution, using a quantitative approach to map real-world data onto the model’s virtual world, and increasing the number of possible risk magnitudes for wind, surge, and rain risk. To assess model realism, we compare FLEE 2.0’s simulated evacuations—specifically its evacuation orders, evacuation rates, and traffic—to available observational data collected during Hurricanes Irma, Dorian, and Ian. FLEE 2.0’s evacuation response is encouraging, given that FLEE 2.0 responds reasonably and differently to all three different types of forecast scenarios. FLEE 2.0 well represents the spatial distribution of observed evacuation rates, and relative to a lower spatial resolution version of the model, FLEE 2.0 better captures sharp gradients in evacuation behaviors across the coastlines and metropolitan areas. Quantitatively evaluating FLEE 2.0’s evacuation rates during Irma establishes model errors, uncertainties, and opportunities for improvement. In summary, this paper increases our confidence in FLEE 2.0, develops a framework for evaluating and improving these types of models, and sets the stage for additional analyses to quantify the impacts of forecast track, intensity, and other positional errors on evacuation. Significance Statement This paper describes and evaluates an updated version of a modeling system [the Forecasting Laboratory for Exploring the Evacuation-system, version 2.0 (FLEE 2.0)] designed to explore relationships between hurricane forecasts and evacuation impacts. FLEE 2.0’s simulated evacuations compare favorably with different types of observational evacuation data collected during Hurricanes Irma, Dorian, and Ian. A statistical comparison with Irma’s observed evacuation rates highlights uncertainties and opportunities for improvement in FLEE 2.0. In summary, this paper increases our confidence in FLEE 2.0, develops a framework for evaluating these types of models, and provides a foundation for additional work using FLEE 2.0 as a research tool.

This study presents a hurricane evacuation traffic simulation framework for the New Orleans-Metairie Metropolitan area, aimed at creating a digital twin for hurricane evacuation planning and operation. The framework integrates signal timing plans and drivers’ hurricane evacuation route choice behavior into the simulation. Initially, the regional road network and local signal timing plans were refined to create a realistic traffic simulation environment. Locations with existing loop detectors along major evacuation routes in the real world with available data during the study period received more attention for result comparison purposes. Subsequently, this study explored a set of scenarios with respect to route choices and the percentage of informed drivers to determine the best parameter set that describes the actual route choices of drivers by comparing simulated traffic with observed traffic during Hurricane Ida (2021). A common observation from all simulation outputs indicated that the evacuation traffic peak occurred around 22 h before the storm landfall. The route choice behavior scenario testing results suggest that a stochastic shortest path model, which minimizes travel time and assumes 70% of drivers are informed, best matches the actual traffic observed at the selected locations during Hurricane Ida. This finding implies that over half of the drivers are aware of travel conditions and prioritize travel time during the evacuation. The simulation results highlight the significance of accurately representing real-world conditions, accounting for drivers’ evacuation route choices, and enhancing simulation outputs by incorporating real-time background traffic loading.

Reinforcement learning for optimizing hurricane evacuation decisions: Hurricane Irma case study

Background/Objectives: Hurricane exposure is a growing public health concern that frequently results in posttraumatic stress symptoms (PTSS) in families. Research suggests that contextual factors, including whether or not individuals evacuate, evacuation stress, perceived sense of control, and peritraumatic distress, contribute to PTSS development. Yet, no known research has evaluated how these variables relate to one another, limiting understanding of how and why evacuation-related circumstances impact PTSS. This study investigated how evacuation experiences and PTSS differ between hurricane evacuees and non-evacuees. Methods: Parents (N = 211) reported on their evacuation experiences and perceptions, as well as their and their child’s PTSS, following Hurricane Ian. Results: Evacuated participants reported greater evacuation stress and greater PTSS in themselves and their child relative to non-evacuated participants. Parents’ sense of control was negatively associated with parent evacuation stress and parent peritraumatic distress in the non-evacuated group only. There were no direct associations between parents’ sense of control and parent or child PTSS in either group. In the non-evacuated group, parent evacuation stress was indirectly related to parent PTSS via parents’ sense of control and parent peritraumatic distress. Similarly, parent evacuation stress was indirectly related to child PTSS via each of the aforementioned variables and parent PTSS in the non-evacuated group only. Conclusions: Stress associated with hurricane evacuation may impact parent’s perceived sense of control, which may contribute to greater parent peritraumatic stress, resulting in greater PTSS among parents and children within families that did not evacuate prior to a hurricane. Findings highlight mechanisms that may inform treatment interventions and public health policy.

Identifying Critical Locations for Traffic Monitoring Devices during Hurricane Evacuations

Local emergency management agencies’ hurricane evacuation decision making and population warning

Extreme natural disasters often lead to large-scale evacuations, and the location of shelters directly impacts the efficiency of these evacuations. Therefore, optimizing shelter location planning is crucial. However, shelter site selection faces numerous challenges, especially when considering the uncertainty of shelter availability. Due to the destructive impacts of disasters, shelters may be damaged, affecting their availability. To address this challenge, this article proposes a data-driven shelter location model to optimize site selection under uncertain availability, thereby improving the efficiency and reliability of emergency evacuation systems. Specifically, the study focuses on shelter locations in hurricane disasters. First, using anomalous meteorological data generated by hurricanes and information on the structural integrity of shelters, a Deep Feedforward Neural Network is trained to predict the availability of shelters. Subsequently, a Wasserstein ambiguity set is constructed based on the model’s prediction results, and a Distributionally Robust Optimization model is established. In terms of the solution method, an affine decision rule and linear relaxation technique are employed to obtain an approximate mixed-integer linear programming (MILP) reformulation. Additionally, a constraint generation algorithm is designed to solve this MILP problem more efficiently, with various acceleration strategies provided to enhance the algorithm’s convergence speed. Numerical experiments demonstrate the effectiveness of the proposed data-driven model and solution algorithm.

The threat of novel pathogens and natural hazards is increasing as global temperatures warm, leading to more frequent and severe occurrences of infectious disease outbreaks and major hurricanes. The COVID-19 pandemic amplified the need to examine how risk perceptions related to hurricane evacuations shift when vaccines become available. This study explores individuals' expected evacuation plans during the early stages of COVID-19 vaccine availability. In March 2021, an online survey was disseminated in Puerto Rico and the US Virgin Islands. An overwhelming majority (72.6%) of respondents said that their vaccination status would not affect their hurricane evacuation intentions. The unvaccinated were significantly more likely to consider evacuating during a hurricane than the vaccinated. Even with vaccines available, respondents suggested they were less likely to evacuate to a shelter during the 2021 season than prior to the COVID-19 pandemic. Respondents generally believed that the risk of contracting COVID-19 at a shelter was greater than the risk of sheltering-in-place during a hurricane. Government officials need to develop and communicate clear information regarding evacuation orders for municipalities that may be more impacted than others based on the trajectory of the storm, social determinants of health, and other factors like living in a flood zone.

<div>Hurricane evacuations generate high traffic demand with increased crash risk. To mitigate such risk, transportation agencies can adopt high-resolution vehicle data to predict real-time crash risks. Previous crash risk prediction models mainly used limited infrastructure sensor data without covering many road segments. In this article, we present methods to determine potential crash risks during hurricane evacuation from an emerging alternative data source known as connected vehicle data that contain vehicle speed and acceleration information collected at a high frequency (mean = 14.32, standard deviation = 6.82 s). The dataset was extracted from a database of connected vehicle data for the evacuation period of Hurricane Ida on Interstate-10 in Louisiana. Five machine learning models were trained considering weather features and different traffic characteristics extracted from the connected vehicle data. The results indicate that the Gaussian process boosting and extreme gradient boosting models outperform (recall = 0.91) other models. Such real-time crash prediction models, leveraging connected vehicle data, could enable transportation agencies to implement proactive countermeasures, such as dynamic speed limit or lane management, during emergency evacuations, thereby enhancing road safety.</div>

Hurricane evacuation analysis with large-scale mobile device location data during hurricane Ian

Deploying Scalable Traffic Prediction Models for Efficient Management in Real-World Large Transportation Networks During Hurricane Evacuations

Understanding hurricane evacuation behavior from Facebook data

As hurricanes become more frequent and destructive, understanding evacuation decision-making is crucial to refining disaster response strategies. Several studies have explored how socioeconomic characteristics such as income and race impact evacuation behavior. Most of these studies focus on a single hurricane event with the geographic extent limited to one or two states, each using a distinct study design, making them difficult to compare. This raises the question of whether findings from isolated cases can be generalized across different hurricane scenarios and geographical settings under a similar study design and consistent parameter definitions. To address this gap, in this study, we conduct a comparative analysis to understand income and racial disparity across multiple hurricane events. The results indicate that, even with a consistent study design, disparities in evacuation among different socioeconomic groups vary on a case-by-case basis. Furthermore, we show that the study design significantly impacts the observed trends within a single case. This highlights the importance of avoiding generalized conclusions based on limited case studies. It further emphasizes how flawed study designs may fail to capture the complexities of real-world behaviors, thereby leading to suboptimal or ineffective policy recommendations or designs.

Using multiple hurricane surveys collected from different areas in the U.S., this study estimates how much distance and how much time were traveled by evacuees to reach safer destinations during a hurricane event. Regression results indicate that flood risk, respondents’ age, income, and education levels are correlated with both the distance and travel time of hurricane evacuation trips. Moreover, to gain deeper insights, we estimated the price and income elasticities of hurricane evacuation trip characteristics. The estimated elasticities of travel distance reveal that travel distance is a necessary and ordinary good, implying that safety is essential and less responsive to price changes. In addition, the estimated elasticities of travel time suggest that travel time is an inferior good, indicating that as income goes up, people tend to spend less time traveling for evacuation. This finding provides logistic implications for emergency management agencies to analyze the evacuation travel demand and ensure safety in vulnerable communities.

Data-Driven Modeling of Hurricane Evacuee’s Individual Decision-Making for Enhanced Hurricane Evacuation Planning: Florida Case Study in the COVID-19 Pandemic

Given the destructive nature of hurricanes in tropical regions, pre-disaster evacuation has emerged as a critical approach for hurricane preparedness. Nevertheless, the compounding effects of natural hazards and the outbreak of infectious diseases, such as Covid-19, significantly challenge hurricane evacuation management. To investigate emergency responses under compound hazards, this study develops an activity-based model to measure the evacuation behaviors of individuals, using Hurricane Irma as a case study. Four scenarios are designed, including a single hurricane hazard, Hurricane Irma compounded with a pandemic like Covid-19, Hurricane Irma compounded with flood damage to the transportation network, and a combination of all these hazards. The metropolis-hasting algorithm is utilized to generate a population with socioeconomic attributes, which is then allocated to census block groups covering Palm Beach, Broward, Miami-Dade, and Monroe Counties in Florida. Datasets from multiple sources are used to measure evacuation decisions, which are subsequently simulated using MATSim. The results highlight the potential impacts of compound hazards on transportation systems, including increased congestions in scenarios involving compounded hurricanes and floods, especially between 10 a.m. and 7p.m. Moreover, a higher proportion of socially vulnerable populations is observed in scenarios involving compounded hurricanes and pandemics, particularly in the Key West area. The developed model could be further applied to measure the indirect impacts of natural hazards on transportation systems.

Abstract In this article, we study the integrated hurricane relief logistics and evacuation planning (IHRLEP) problem, integrating hurricane evacuation and relief item pre‐positioning operations that are typically treated separately. We propose a fully adaptive multistage stochastic programming (MSSP) model and solution approaches based on two‐stage stochastic programming (2SSP). Utilizing historical forecast errors modeled using the auto‐regressive model of order one, we generate hurricane scenarios and approximate the hurricane process as a Markov chain, and each Markovian state is characterized by the hurricane's location and intensity attributes. We conduct a comprehensive numerical experiment based on case studies motivated by Hurricane Florence and Hurricane Ian. Through the computational results, we demonstrate the value of fully adaptive policies given by the MSSP model over static ones given by the 2SSP model in terms of the out‐of‐sample performance. By conducting an extensive sensitivity analysis, we offer insights into how the value of fully adaptive policies varies in comparison to static ones with key problem parameters.

Risk theories and empirical research indicate that a variety of factors can influence people’s protective decisions for natural hazards. Using data from an online survey that presented coastal U.S. residents with a hypothetical hurricane scenario, this study investigates the relative importance of cognitive risk perceptions, negative affect, efficacy beliefs, and other factors in explaining people’s anticipated evacuation decisions. The analysis finds that multiple factors, including individual and household characteristics, previous experiences, cognitive and affective risk perceptions, and efficacy beliefs, can help predict hurricane evacuation intentions. However, the largest amount of variance in survey participants’ evacuation intentions is explained by their evacuation-related response efficacy (coping appraisals) and their perceived likelihood of getting hurt if they stay home during the storm. Additional analysis explores how risk perceptions and efficacy beliefs interact to influence people’s responses to risk information. Although further investigation in additional situations is needed, these results suggest that persuading people at high risk that evacuating is likely to reduce harm can serve as an important risk communication lever for motivating hurricane evacuation.

Navigating dual hazards: Managing hurricane evacuation and sheltering operations Amidst COVID-19

Information retrieval and classification of real-time multi-source hurricane evacuation notices