Timely hazard assessments are needed to assess post-fire debris flows that may
impact communities located within and adjacent to recently burned areas. Implementing
existing models for debris-flow probability and magnitude can be time-consuming because
the geographic extent for applying the models is manually defined. In this study, a model is
presented for predicting locations of post-fire debris-flow erosion. This model is further
calibrated to identify the geographic extent for applying post-fire hazard assessment
models. Aerial photographs were used to map locations of post-fire debris-flow erosion and
deposition in the San Gabriel Mountains. Terrain, burn severity, and soil characteristics
expected to influence debris-flow erosion and deposition were calculated for each mapped
location using 10-m resolution DEMs, GIS data for burn severity, and soil surveys.
Multiple logistic regression was used to develop a model that predicts the probability of
erosion as a function of channel slope, planform curvature, and the length of the longest
upstream flow path. The model was validated using an independent database of mapped
locations of debris-flow erosion and deposition and found to make accurate and precise
predictions. The model was further calibrated by identifying the average percentage of the
drainage network classified as erosion for mapped locations where debris flows transitioned
from eroding to depositing material. The calibrated model provides critical information
for consistent and timely application of post-fire debris-flow hazard assessment
models and the ability to identify locations of post-fire debris-flow erosion.
