Climate risksRegulatory hazards

Climate risks — Regulatory hazards

The 28 physical climate hazards of the EU Taxonomy (Annex I, Appendix A), line by line, with what Darwin's method models for each — quantified, partial or not covered — and what it measures instead where the correspondence does not hold.

Climate risks — Regulatory hazards

Darwin prices eleven climate hazards. The list a report is usually held to is a different one: Appendix A of the EU Taxonomy Climate Delegated Act (Annex I), the only normative European list of physical climate hazards, which the CSRD (ESRS E1) reuses as is. It is what clients mean by "the ~20 climate indicators", and it holds 28 hazards in four families — temperature, wind, water, solid mass — each split between chronic (a slow drift, with no return period) and acute (a dated event).

The Resilience page carries that list in its Regulatory hazards tab. This page reproduces it, with the status Darwin's method carries on each line.

The two lists are not in bijection, in either direction. One regulatory line can be answered by several Darwin hazards — heavy precipitation by both extreme precipitation and hail. One Darwin risk can answer a line it does not name — soil erosion is answered on the nature side. And a Darwin hazard can sit outside the list entirely: earthquake is geophysical, not climatic, and is shown beside the list so its count stays exact. Reading either list as the other is the mistake the tab exists to prevent.

Two statuses, never merged

The tab shows two columns that answer two different questions.

  • Darwin method — what the methodology models at all. A property of Darwin, identical on every deployment, and the column reproduced below.
    • Quantified — priced in euros per site, property damage and business interruption kept apart.
    • Partial — modelled, but not readable as this line on its own: either it is a parameter of another indicator, or Darwin answers a narrower question than the line asks.
    • Not covered — not modelled. Shown rather than hidden: a list claiming to cover the line would be false.
  • Data on this environment — whether the environment you are working in holds the reference data behind the Darwin hazards answering the line: available, partly available (a line answered by several hazards, only some of which have data), or no data. A line can be quantified in method and have no data in fact; the row then says modelled by Darwin, but not measurable here rather than choosing one of the two. A line that names no Darwin hazard carries no data status, because there is nothing to hold data for. This column depends on the deployment, so it is not reproduced here.

On the current list: 8 quantified, 5 partial, 15 not covered.

Temperature

Regulatory hazardRegimeDarwin methodAnswered byWhat to read
Changing temperature (air, freshwater, marine water)ChronicNot coveredOnly the extremes are priced; the drift of the mean is not.
Heat stressChronicQuantifiedExtreme heatWBGT turned into a loss of work capacity (ILO / Kjellström), modulated by work intensity and air conditioning. Business interruption only.
Temperature variabilityChronicNot covered
Permafrost thawingChronicNot covered
Heat waveAcuteNot coveredDarwin measures chronic heat stress on the workforce, not a heat wave — a threshold event (equipment failure, cold-chain break, production stop) with a return period. Reading the two as one is the most likely way to overstate the coverage.
Cold wave/frostAcuteNot covered
WildfireAcuteQuantifiedWildfireFire Weather Index converted into fireline intensity (Van Wagner), then into a loss rate by distance to the wildland-urban interface.

Wind

Regulatory hazardRegimeDarwin methodAnswered byWhat to read
Changing wind patternsChronicNot covered
Cyclone, hurricane, typhoonAcuteQuantifiedCycloneSTORM wind fields converted to sustained wind, Emanuel (2011) damage curve.
Storm (including blizzards, dust, and sandstorms)AcuteNot coveredDarwin covers the tropical cyclone only; the extratropical storm is a distinct hazard field and is not modelled.
TornadoAcuteNot covered

Water

Regulatory hazardRegimeDarwin methodAnswered byWhat to read
Changing precipitation patterns and types (rain, hail, snow/ice)ChronicPartialDarwin carries the signal without an indicator of its own: climate change factors deform the flood and drought curves, so the shift is priced, but it cannot be read as a separate line.
Precipitation or hydrological variabilityChronicPartialSame situation as changing precipitation patterns.
Ocean acidificationChronicNot coveredBelongs to the nature side rather than the physical-loss engine.
Saline intrusionChronicNot covered
Sea level riseChronicPartialFloodSea level is a parameter of the coastal flood arm, not an indicator: it is folded into the water depth at every horizon, precisely to avoid counting it twice, so it is really taken into account but not reportable on its own.
Water stressChronicQuantifiedDroughtAqueduct baseline water-stress score turned into a production loss, modulated by the site's water dependence and, when volumes are stated, by a quantitative criticality.
DroughtAcuteNot coveredDarwin measures chronic water stress, not an acute drought — a dated event with a return period. The Aqueduct score carries no drought index, no return period and no event; reading the two as one overstates the coverage.
Heavy precipitation (rain, hail, snow/ice)AcuteQuantifiedExtreme precipitation · HailTwo Darwin hazards answer this single line: rainfall above the local drainage design (EN 752), and hail by maximum hailstone diameter (Schmid 2024).
Flood (coastal, fluvial, pluvial, ground water)AcuteQuantifiedFloodFluvial arm (GIRI) and coastal arm (Deltares), protection standards (FLOPROS), JRC depth–damage curves.
Glacial lake outburstAcuteNot covered

Solid mass

Regulatory hazardRegimeDarwin methodAnswered byWhat to read
Coastal erosionChronicNot covered
Soil degradationChronicPartialClay shrink-swellDarwin treats clay shrink-swell only. The European meaning — chemical, biological and physical degradation — is far broader, so the correspondence is partial at best.
Soil erosionChronicPartialSoil and sediment retention (nature)Priced on the nature side, through the ecosystem service, not valued through the physical-loss engine. It sits in this climate list because Appendix A lists it, and the correspondence is stated rather than hidden.
SolifluctionChronicNot covered
AvalancheAcuteNot covered
LandslideAcuteQuantifiedLandslideNASA susceptibility crossed with an extreme-precipitation trigger; the class→loss shape is calibrated on the NASA Global Landslide Catalog.
SubsidenceAcuteQuantifiedSubsidenceHerrera-García global susceptibility, treated as a chronic creep although the regulation lists it as acute.

Beyond the regulatory list

Earthquake is priced by Darwin — GSHAP peak ground acceleration on HAZUS fragility curves — and absent from Appendix A, which lists climate hazards only. It is shown apart from the 28 so the count of the regulatory list stays exact.

"Quantified" is a statement about the method, not a calibration grade. Four of the eight quantified lines rest wholly or partly on hazards Darwin tags proposal — drought (water stress), extreme precipitation (heavy precipitation, beside hail), landslide and subsidence — and those figures sit outside the totals by default. The tag is on the climate risks page; read the two together.

What is not published yet

Appendix A is a table of labels with no definitions: nothing in it separates heat stress from heat wave except the sentence each line carries. Each line has room for the IPCC AR6 climatic impact-driver that defines it and the ETCCDI / WMO reference index that makes it measurable; both are empty on all 28 lines, because authoring them is methodology work in progress rather than a by-product of this list, and the tab says so above the table. Guessing them would be exactly the error those columns exist to prevent. Authoring both columns is a methodology deliverable in progress; it lands with the loss thresholds that also gate the ordinal grade.

Changing a line's Darwin status changes what Darwin claims to cover, so it is a methodology decision, signed off as one — not an edit to the list.