Climate Adaptation KPIs: Measure Resilience Beyond CO₂

Measuring climate resilience means putting numbers on losses, downtime and recovery, before and after you act. That goes beyond CO₂ reduction, and it needs different data: hazard data from public sources, plus your own records of what weather has already cost your sites.

What resilience measurement can and cannot show

Resilience is the ability of a site or a company to absorb a climate impact and return to normal operation. Adaptive capacity is what you can still change before the next event. Neither is a single number. As the OECD notes: “Measuring adaptation progress is complex and cannot be captured in a single metric.”

The core problem is attribution. A working measure shows up as damage that did not happen, and a flood barrier can look useless for 20 years and then pay for itself in one night. So I measure three things: exposure, losses and recovery. Then I compare them against a baseline without the measure, and I label that comparison as modelled. It is also one reason why investors still put more capital into mitigation than into adaptation: the return is harder to show.

Data sources and measurement methods for climate resilience

Public data tells you what a hazard does in a region. Only your own loss, repair, insurance and downtime records tell you what it costs at your sites, so start by collecting those. The public layer then covers the probability of the hazard.

For scale: weather and climate extremes caused an estimated €822 billion of losses in the EU between 1980 and 2024, and less than 20 percent was privately insured (EEA). Official sources do not give company-level exposure classes, so I use my own five-tier threshold scale per hazard, referenced to observations and RCP4.5 and RCP8.5 projections.

A short KPI set per hazard: exposure, damage avoided, downtime

Take the smallest set that covers the hazards you found: one economic, one operational and one coverage indicator is usually enough. Each KPI needs a formula, its data and a baseline.

Downtime days and recovery time are the same figures a business continuity plan tracks, so the data often exists already. To build the set, work in this order:

1. Set the scope: sites, assets and time horizon.
2. Establish which hazards reach those assets, and how strongly.
3. Pick the indicators from the table above.
4. Fix a baseline from existing data, then a target with a date.
5. Decide who produces the evidence and who checks it, before the first measurement.

To judge your adaptive capacity, answer six questions per critical site: Is there backup power? Is there an alternative site or supplier? Can critical equipment be moved or raised? Is the damage insured, and up to which amount? Is there a written plan for the event, and was it ever tested? Who decides when it is triggered? Each “no” is a gap you can score.

Worked example: expected annual loss before and after one adaptation measure

The following is an illustrative calculation with assumed figures, not a client case. One logistics site, one hazard (flooding from heavy rain), three event classes, and a measure of flood barriers plus raised electrics for €150,000 with a lifetime of 20 years.

The EAL is the sum of probability × loss: 0.08 × 100,000 + 0.015 × 600,000 + 0.005 × 1,500,000 = €24,500. After the measure it is €8,000, so the avoided loss is €16,500 per year. At a discount rate of 3 percent over 20 years, the present value is about €245,000, and the BCR is 245,000 ÷ 150,000 ≈ 1.6. The result is only as good as the damage ratios, so ask your insurer or an engineer to check them. And the probabilities are not constant: if heavy rain gets more frequent, the real BCR is higher than this one.

Where adaptation KPIs land in reporting: ESRS E1-9 and the voluntary VSME C4 module

Under the ESRS, disclosure E1-9 covers the anticipated financial effects of material physical risks, where the risk is material, so EAL and the share of assets in high-hazard locations are the figures it asks for. The Voluntary Standard (VS), known as VSME, is voluntary. Its comprehensive module has a field C4 for climate risks. Nobody is legally required to do a resilience assessment for it, and anyone who says otherwise is selling something. The EU Taxonomy asks for a climate risk and vulnerability assessment only if you report or finance an activity as aligned. See the ISO 14091 guide for the method behind it.

Adaptation KPIs also enter loan documentation. In sustainability-linked loans, a covenant-grade KPI needs a written calculation method before signing, a baseline from data that existed at signing, and evidence a third party can check. Share of assets assessed, retention area and recovery time from an incident log clear that bar. Modelled figures such as avoided losses usually do not, because the comparison case never happened.

In a climate risk analysis for a dealer group in southern Germany, a few critical sites carried a disproportionate share of the physical risk, mainly hail and heavy rain. That concentration is the number worth showing a lender, not the portfolio average. For the method, see physical and transition risks.

FAQs