Scientific framing¶
Post-fire debris flows are fast, destructive mixtures of water and sediment initiated on recently burned steep terrain when short-duration, high-intensity rainfall mobilizes loose material. Wildfire is the conditioning factor: it removes protective vegetation, induces soil water repellency (hydrophobicity), and lowers root cohesion, so that storms far smaller than ordinary landslide-triggering events can generate a flow. This makes them a special case of landsliding in which burn severity (e.g. dNBR) and fire-altered soil properties are first-order inputs — the feature that separates this hazard from the core GAIA shallow/deep landslide susceptibility, which is rainfall- and groundwater-driven and carries no fire term.
State variables & observables¶
(Burn severity / dNBR, soil water repellency, short-duration rainfall intensity thresholds, sediment supply; seismic/infrasound detection, post-event DEM/lidar differencing for run-out and volume.)
Data — what we ingest¶
(Link to DataHub: burn severity (MTBS/BAER), high-resolution rainfall, terrain, and the soil layers of Pillar 1. This is the one landslide hazard whose data stack requires a wildfire layer.)
Models¶
(Link to ModelHub: the Landlab LandslideProbability workflow run with
fire-reduced cohesion and burn severity (see
Pillar 2 §2.5), rainfall intensity–duration thresholds,
run-out modeling, and multi-sensor detection.)
Evaluation & metrics¶
(Link to HazEvalHub: triggering detection, run-out extent agreement, warning lead time.)
Connection to use cases¶
The driving use case is the 2025 Stehekin post-fire debris flow; the
modeling pipeline lives in
gaia-hazlab/landlab-debrisflow.