Brief Description:

Due to the high spatial variability and uncertainty in input parameters at regional scales, implementing mechanistic landslide predictive models often requires many assumptions that might affect the reliability of the predictions. Understanding how changes in these inputs affect CRISIS's outcomes is key to evaluating the sensitivity and reliability of the results.

We conducted a sensitivity analysis for CRISIS's key input parameters and features.

Key Takeaway:

Model parameters and features were ranked from most to least critical based on their influence on the predicted cumulative landslide area density as well as characteristics of individual landslides.

This ranking helps prioritize which parameters and features should receive higher-quality and higher-resolution data in future predictions, while allowing less critical inputs to be simplified without significantly affecting results, thereby reducing computational costs