Combining release and runout in statistical landslide susceptibility modeling.

Area of Science:

• Geosciences
• Geomorphology
• Natural Hazards

Background:

• Landslide susceptibility modeling often separates release and runout phases.
• Integrating these phases is crucial for accurate hazard assessment.

Purpose of the Study:

• To introduce and compare two novel GIS-based approaches for combining landslide release and runout.
• To evaluate their performance in predicting landslide susceptibility and impact areas.

Main Methods:

• Employed a constrained random walk for downslope routing of mass points.
• Utilized probability density functions (PDF) and cumulative density functions (CDF) of reach angles and travel distances.
• Compared a bottom-up approach (A) and a top-down approach (B).

Main Results:

• Approach A (bottom-up) yielded quantitative spatial probability but lost release susceptibility signals.
• Approach B (top-down) retained release signals, performed better, and reproduced observed impact areas more effectively.
• Model performance and conservativeness were sensitive to the choice of PDF/CDF parameters (reach angle, travel distance).

Conclusions:

• The top-down approach (B) is superior for integrating landslide release and runout in GIS-based modeling.
• Accurate landslide hazard assessment requires considering both release and runout dynamics.
• Careful selection of statistical functions (PDF/CDF) is critical for robust modeling outcomes.