What it measures

Nine domains, their weights, and which of them are scored against a scientific boundary.

The nine domains

DomainWhat it measuresWeightBasis
Land coverForest cover and canopy vitality0.25boundary
WildfireActive hotspot burden0.20contextual proxy
AirParticulate and nitrogen dioxide burden0.15boundary
Ocean warmingSea surface temperature anomaly0.10boundary
Ocean acidificationSurface aragonite saturation0.10boundary
CryosphereGlacier and sea ice signal0.10contextual proxy
BiodiversitySpecies pressure and habitat integrity0.10boundary
AnthroposphereHuman modification and density pressure0.10contextual proxy
ConservationProtected area coverage0.08contextual proxy

A boundary domain is scored as distance from a scientific threshold. A contextual proxy is a signal that tracks the same story usefully without being a boundary control variable, and it is labelled that way so nobody reads it as one.

Why the weights do not total 1

They total 1.18, and that is deliberate. The score renormalizes by the weight it actually covered, so a region that reports six domains is scored on those six rather than being penalised for the three it could not report. What drops instead is the confidence, which is the subject of the next page.

How the global number is formed

Regional scores are combined by population exposure rather than by land area, so a domain's contribution reflects how many people it reaches. The planetary trend arrow is derived from the cryosphere domain alone. That is a real narrowing and it is stated plainly on limits.

What is not on this page

The normalizers, the per-domain boundary thresholds, the protected area saturation curve and the humility ceiling that keeps any signal from reading as a solved 100 all exist in writing. They are not published at this site, so this page states the shape of the method and not its constants. See limits for what that costs a reader.