Resilience
How Darwin prices the risk a site carries: two families — climate hazards striking the asset, and nature dependencies failing upstream — read on the same money terms, at site, company and portfolio level.
Resilience
Resilience is the module that prices the risk a site carries. It answers one question — how much money is at risk at this location, today and under future scenarios? — and it answers it from two directions at once.
Two risk families
| Family | What it prices | Drivers |
|---|---|---|
| Climate | A hazard striking the asset: water in the building, wind tearing the roof off, ground moving under the foundations. | 14 hazards |
| Nature | A loss travelling through an ecosystem service the site relies on: pollination failing, water no longer purified upstream, soil no longer held. | 9 final ecosystem services |
Twenty-three risk drivers in total, on one screen and in the same money terms, so they can be read, ranked and aggregated together.
The climate list is Darwin's own. The list a report is usually held to is the EU Taxonomy's — 28 climate hazards in Annex I, Appendix A — and the module answers it line by line, including the lines it does not cover: see Regulatory hazards.
Why earthquake sits in the climate family. The climate family is the direct physical family: a hazard reaching the asset. Some of its members — earthquake, and the ground-movement trio — are geophysical and have no climate trigger at all, so no scenario moves them. They are grouped with the climate hazards because they are priced exactly the same way, through a vulnerability curve on a hazard intensity, and splitting off a third family for them would buy nothing. What it does mean is that their change against the baseline is zero by construction, which is why the module reports a level alongside every delta.
Two loss channels, never summed
Both families report on the same two channels, and the two are always kept apart:
- Property damage — a share of the site's asset value: what is physically destroyed. A stock.
- Business interruption — a share of one year's revenue: output lost while the site is stopped or throttled. A flow.
Adding a stock to a flow mixes units. Any combination happens downstream, never here.
Only the climate family produces property damage — a dependency failing upstream does not damage the building. On the revenue channel the two families can meet, and the catalogue is built so that they meet without pricing the same loss twice.
How the two families are reconciled
The module carries 23 risk drivers — the 14 climate hazards plus the 9 SEEA-EA final ecosystem services. The nature side is re-keyed onto the service (never the ecosystem component): a site depends on the service the components deliver, not on the raster input.
The two families are kept apart by the catalogue itself, not by a rule applied afterwards. A service that was only ever the nature-side reading of a hazard the climate family already prices is simply not in the catalogue. Water supply read the same Aqueduct field as the drought curve; flood mitigation, storm mitigation and local climate regulation had no risk driver of their own — the only input pointing at each of them was the corresponding hazard. All four were removed rather than arbitrated. When it really is the same data, one of the two goes.
What remains is the opposite case: two measurements of two genuinely distinct processes, which coexist. The four ground-and-rain hazards (landslide, clay shrink-swell, subsidence, extreme precipitation) each carry a business-interruption channel sitting next to a nature axis without duplicating it — the climate family prices a discrete event (the slope lets go, the site stops for the works), the dependency path prices the chronic degradation of a service (erosion, siltation, soil loss), and the two read entirely disjoint inputs. Soil and sediment retention stays in the catalogue for exactly that reason, though it shares an ENCORE reference with landslide: sharing a reference is not being the same risk.
An arbitration rule remains in the engine for the day a genuine overlap reappears — two estimates of the same revenue loss are arbitrated, never summed, and the climate estimate wins, a loss priced through a calibrated vulnerability curve taking precedence over one reached by a dependency proxy, with the nature estimate retained as corroboration. Today it fires on nothing: every hazard and every nature axis lands on a distinct risk. On the asset-value channel the question does not arise — only the climate family produces property damage.
Splitting the figures by family. A climate report (TCFD-style) wants the climate figures on their own; a nature report (TNFD-style) wants the nature figures on their own. In Resilience this split is clean: every risk belongs to one family and one only, so the climate total and the nature total add up to the module total. Nothing is counted twice and nothing is dropped. When the engine holds two estimates of the same loss, one from each side, it keeps one and notes the other as corroboration — but that is a note on the estimate, not a second family on the risk, and the export always carries the risk's family.
Outside Resilience, the picture is different because the unit is not a risk but a row — a commodity from a given country on the Sourcing page, a site on the Exposure page. A row is listed in the climate view as soon as one of its drivers is a climate hazard, and in the nature view as soon as one of them is a nature dependency. A row that has both — Brazilian soy facing both drought and pollinator decline, say — is therefore listed in both views. Adding up the two views then gives more rows than there are in total. This is expected, and the interface shows how many rows sit on both sides, so the reader sees the overlap instead of having to work it out.
Where to start
New to physical and nature risk? Each family has a jargon-free primer: climate and nature. The full methods are on the family overview pages, the damage functions show the curve behind every climate figure, and regulatory hazards maps the climate family onto the EU Taxonomy list.