Mapa de Riesgo de Incendios Forestales (AHP 4 factores)

Riesgosv01.50 créditos/per_km2

Responsable:Niccolò Calandri

Suscettibilità strutturale agli incendi boschivi a 30m. Sovrapposizione AHP pesata di 4 fattori: peso fuel ESA WorldCover (0.35), NDVI estivo invertito (0.25), Indice di Aridità Globale invertito (0.20), LST estiva Landsat-8/9 (0.20). Rinormalizzazione NaN-aware per pixel. Output: raster a 2 bande (score 0-1, qualità 0-1). Composto interamente lato server su GEE.

Fuentes de Datos y Referencias

Fuentes

ESA_WORLDCOVER_V200

COPERNICUS_S2_HARMONIZED

GLOBAL_AI_YEARLY

LANDSAT_LC08_C02_T1_L2

LANDSAT_LC09_C02_T1_L2

Metodología

Wildfire Risk Map

Formula

score(p) = Σᵢ wᵢ·fᵢ(p)·validᵢ(p) / Σᵢ wᵢ·validᵢ(p)

Factors

Factor	Source	Weight	Normalization
Vegetation type	ESA/WorldCover/v200	0.35	Lookup table (Scott & Burgan Med fuel models)
Summer NDVI	COPERNICUS/S2_SR_HARMONIZED	0.25	1 - clip((ndvi-0.1)/0.6, 0, 1)
Aridity Index	global_ai_yearly / 10000	0.20	1 - clip(ai/0.65, 0, 1)
Summer LST	LANDSAT/LC08+LC09 ST_B10	0.20	clip((lst_C-20)/20, 0, 1)

Calibration

Class thresholds [0.20, 0.35, 0.50, 0.70] calibrated against EFFIS GlobFire 2020-2024 burned area on 10 Italian benchmark polygons (Phase 4).

References

Oliveira, S. et al. (2021). Natural Hazards, AHP+GIS Iberian Mediterranean.
Conedera, M. et al. (2018). Int. J. Wildland Fire, Italy national assessment.
Tonini, M. et al. (2020). Fire 3(3) 43, RF/MaxEnt/LR comparison.
Santangelo, N. et al. (2022). Remote Sensing 14(19) 4803, XGBoost+SHAP Sardinia.
San-Miguel-Ayanz, J. et al. (2013). EFFIS methodology, JRC.