AVL · Machine learning
Avalanche
XGBoost · spatially blocked CV · hold-out AUC 0.99 · v1.1 · 2026-04
French & Italian Alps — Chamonix corridor at peak
Machine-learning susceptibility from terrain and snow-climate, trained on observed avalanche events.
What ships
Surfaces — calibrated 0–1 probability per pixel.
Every release: 0–1 score · 5-class band · per-pixel uncertainty · input-coverage mask · STAC item · methodology doc.
Inputs
DEM slope/aspect, mean snowfall, winter temperature, terrain ruggedness
Validation
Spatially blocked hold-out AUC 0.99 — cross-validation blocks prevent leakage across nearby pixels.
Version history
What changed between releases — the same notes that ship with the data. Newest first.
v1.1
2026-04
current
- Compact uint16 encoding. The surfaces were re-encoded from float32 to compact integer COGs with a documented scale (physical = raw × scale, nodata 65535) — dramatically smaller and self-describing, with no loss of precision at the delivered resolution. Method and values are unchanged from v1.0.
v1.0
2026-04
- Initial release. Machine-learning avalanche susceptibility (gradient-boosted trees, XGBoost) from terrain and snow-climate — DEM slope/aspect, mean snowfall, winter temperature, and terrain ruggedness — trained on observed avalanche events with spatially-blocked cross-validation (hold-out AUC 0.99). Ships a five-class susceptibility map plus per-pixel uncertainty and coverage.
Other hazards
LandslideWildfireEarthquakeWindstormExtreme heatExtreme coldFrost / freezeLightningSoil erosionHurricaneDrought