Seismic multi-hazard and impact estimation via causal inference from satellite imagery

Area of Science:

• Geophysics
• Remote Sensing
• Disaster Management

Background:

• Accurate post-earthquake assessment is crucial for emergency response and rehabilitation.
• Existing methods struggle with complex, event-specific causal links in cascading seismic hazards.
• Geospatial data and satellite imagery are abundant but underutilized for integrated hazard analysis.

Purpose of the Study:

• To develop a rapid seismic multi-hazard and impact estimation system.
• To improve the accuracy and resolution of regional-scale hazard and damage assessments.
• To model causal dependencies among earthquake-triggered secondary hazards and building damage.

Main Methods:

• Utilizing advanced statistical causal inference techniques.
• Integrating remote sensing data, specifically satellite images.
• Jointly inferring multiple hazards (landslide, liquefaction) and building damage.

Main Results:

• The system achieves accurate, high-resolution estimations on a regional scale.
• Incorporating causal dependencies significantly enhances estimation accuracy compared to existing systems.
• Quantitative causal mechanisms among multi-hazards and impacts were revealed across diverse seismic events.

Conclusions:

• The developed system offers a novel approach for extracting and utilizing complex interactions of seismic hazards and impacts.
• Improved understanding of seismic geological processes and disaster response effectiveness.
• Establishes a new standard for rapid, integrated seismic risk assessment.