Evac-Cast: An Interpretable Machine-Learning Framework for Evacuation Forecasts Across Hurricanes and Wildfires

TL;DR

Evac-Cast is an interpretable machine learning framework that accurately predicts evacuation rates during hurricanes and wildfires using multi-source data, offering a transparent tool for emergency decision-making without relying on psychosocial surveys.

Contribution

This study introduces Evac-Cast, a novel interpretable ML framework that predicts evacuation rates across hazards using macro-level features, eliminating the need for psychosocial survey data.

Findings

Achieves mean absolute errors of 4.5% for hurricanes and 3.5% for wildfires.

SHAP analysis identifies hazard intensity as the most important feature.

Models perform well without explicit psychosocial variables.

Abstract

Evacuation is critical for disaster safety, yet agencies lack timely, accurate, and transparent tools for evacuation prediction. This study introduces Evac-Cast, an interpretable machine learning framework that predicts tract-level evacuation rates using over 20 features derived from four dimensions: hazard intensity, community vulnerability, evacuation readiness, and built environment. Using an XGBoost model trained on multi-source, large-scale datasets for two hurricanes (Ian 2022, Milton 2024) and two wildfires (Kincade 2019, Palisades--Eaton 2025), Evac-Cast achieves mean absolute errors of 4.5% and 3.5% for hurricane and wildfire events, respectively. SHAP analysis reveals a consistent feature importance hierarchy across hazards, led by hazard intensity. Notably, the models perform well without explicit psychosocial variables, suggesting that macro-level proxies effectively encode behavioral signals traditionally captured through time-consuming surveys. This work offers a survey-free, high-resolution approach for predicting and understanding evacuation in hazard events, which could serve as a data-driven tool to support decision-making in emergency management.