Cross-comparative analysis of evacuation behavior after earthquakes using mobile phone data

Area of Science:

• Disaster Science
• Mobility Studies
• Geophysics

Background:

• Understanding evacuation mobility dynamics after disasters is vital for effective disaster relief.
• Limited empirical evidence exists on individual evacuation behavior across multiple disaster events.

Purpose of the Study:

• To analyze large-scale GPS mobility data to understand individual evacuation behavior following major earthquakes.
• To identify key factors influencing evacuation decisions and distances.

Main Methods:

• Analysis of anonymized GPS trajectories from over 1 million mobile phone users before and after four major Japanese earthquakes.
• Cross-comparative analysis of evacuation patterns across different seismic events and urban characteristics.

Main Results:

• Individual evacuation probability is strongly dependent on experienced seismic intensity, with a critical threshold around seismic intensity 5.5.
• Evacuation probability patterns were consistent across all analyzed earthquakes, regardless of event specifics or urban setting.
• The distance individuals evacuate is independent of the seismic intensity experienced.

Conclusions:

• Evacuation behavior following earthquakes exhibits a predictable pattern primarily driven by seismic intensity.
• These findings can aid in developing efficient computational models for predicting evacuee numbers using seismic intensity and population data.