HEAT · Extreme value
Extreme heat
GEV on ERA5 mx2t + elevation lapse downscale · v1.1.1 · 2026-07
Indo-Gangetic plain to the Tibetan plateau — elevation lapse in action
The 100-year extreme high temperature, from 30 years of ERA5 climate reanalysis, downscaled to 90 m.
What ships
Surfaces + Returns — 0–1 score derived from the 100-year Tmax, plus per-return-period annual-max Tmax (rp5/10/25/50/100, °C). Full-resolution (GSHHG-f) coastline: barrier islands now scored.
Every release: 0–1 score · 5-class band · per-pixel uncertainty · input-coverage mask · STAC item · methodology doc.
Inputs
ERA5 monthly max temperature 1991–2020, DEM for elevation lapse correction
Version history
What changed between releases — the same notes that ship with the data. Newest first.
Surfaces
v1.1.1
2026-07
current
- Full-resolution coastline (GSHHG-f). All four layers are re-clipped against the canonical GSHHG-f shoreline (~100 m native), so barrier islands and narrow coastal strips that previously read as NoData now carry a score. Any query over real land returns a result.
- Model unchanged. Susceptibility, classification, uncertainty, and coverage values are identical to v1.1 (differences are coastline-only); this vintage is a land-mask refresh, not a model revision.
- Known limitation carried forward (see model card): the susceptibility score is compressed toward the high end — most of the warm world reads "high" to "very high," so the surface discriminates well between hot and cold climates but weakly *within* the hot world. A calibration revision is planned for a future vintage.
v1.1
2026-04
- Compact uint16 encoding. The surfaces were re-encoded from float32 to compact integer COGs with a documented scale (physical = raw × scale, nodata 65535) — 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. Susceptibility derived from a Generalized Extreme Value (GEV) fit to 30 years of ERA5 monthly maximum temperature (mx2t, 1991–2020). The 100-year return-period Tmax is ranked to a 0–1 percentile scale, with a five-class map (Very Low polar/alpine → Very High tropical/desert), plus per-pixel uncertainty and coverage.
- Elevation-lapse downscale to 90 m. The coarse ERA5 field is corrected with a standard −6.5 °C/km environmental lapse rate against the DEM, so the sharp temperature drop from lowland plains up onto plateaus and mountain ranges is resolved at 90 m rather than smeared across the native grid.
Returns
v1.0
2026-07
current
- Initial release. Per-return-period annual-maximum daily temperature (Tmax) from a per-cell generalized extreme value (GEV) fit to 30 years of ERA5 (1991–2020) annual-maximum 2 m temperature (mx2t), at the 5-, 10-, 25-, 50-, and 100-year return periods.
- Physical °C on one shared scale + offset. Values are daily-max air temperature in °C; because temperature is signed (annual-max Tmax dips below 0 °C on permanent-ice cells), the uint16 encoding carries both a SCALE and an OFFSET, shared across all five return periods, so raw * scale + offset yields °C and no negative values are clipped.
- Elevation-lapse downscale. The 0.25° return levels are brought to 90 m by adjusting to the terrain height with an environmental lapse rate, so valleys read hotter and ridges cooler than a plain interpolation would give.
- GSHHG-f canonical land mask (~100 m native): every real coast returns a value; ocean and Antarctica are NoData.
- Same fit as Extreme Heat Surfaces. The 100-year Tmax here is exactly the field the Extreme Heat Surfaces line normalizes into its 0–1 susceptibility score.
Other hazards
LandslideAvalancheWildfireEarthquakeWindstormExtreme coldFrost / freezeLightningSoil erosionHurricaneDrought