Incorporating shear stiffness into post-fire debris flow statistical triggering models

Commonly used post-fire debris flow statistical triggering models consider predictor variables that account for; rainfall intensity, rainfall accumulation, area burned, burned intensity, geology, slope, and others. These models represent the physical process of debris flow initiation and subsequent failure by quantifying near-surface soil characteristics. Shear wave velocity as a proxy for sediment shear stiffness informs the likelihood of particle dislocation, contractive or dilative volume changes, and downslope displacement that result from flow-type failures. This broadly available variable common to other hazard predictions, such as liquefaction analysis, provides good coverage in the watersheds of interest for debris flow predictions. A logistic regression is used to compare the new variable against currently used variables for predictive post-fire debris flow triggering models. We find that the new variable produces slightly improved performance in prediction of triggering while better capturing the physics of flow-type failure. Additional suggestions are presented for utilizing statistical cross-validation methods to advance prediction performance and the utility of different variables for quick assessment of likelihood during post-fire rainfall events.

Automated summary

This study compared a new variable with currently used variables in predicting post-fire debris flow triggering models using a logistic regression model. The results showed that the new variable slightly improved the prediction performance and better captured the physics of flow-type failure. The paper also offered suggestions for utilizing statistical cross-validation methods to improve prediction performance and the utility of different variables for quick assessment of likelihood during post-fire rainfall events.