Skip to content
13 min read

Physical risks and transition risks: How to assess climate risks

Storm waves crashing against a pier with a lighthouse

Physical climate risk is the direct financial impact of climate hazards on a business, floods, heat waves, drought, storms. Transition risk is the financial impact of the response to those hazards: new climate policy, new technology, shifting markets, reputational pressure. Both trace back to the same variable, greenhouse gas emissions, from opposite directions. The CSRD expects companies in scope, through ESRS E1, to use climate scenario analysis for both risk types: at least one high-emissions scenario for physical risks and a 1.5°C pathway with no or limited overshoot for transition risks.

Live data: See how physical climate hazards such as flooding, drought and heat hit specific locations in the Fiegenbaum Atlas climate risk dashboard.

Physical vs transition climate risk: the difference in one table

Physical risk is the direct impact of climate hazards on assets, operations and supply chains. It splits into acute risk, single events such as floods, storms or heat waves, and chronic risk, gradual shifts such as rising average temperatures, changing rainfall or sea-level rise. Transition risk is the impact of the move to a low-carbon economy: policy and legal risk (new regulation, carbon pricing), technology risk (assets or processes becoming obsolete), market risk (shifting customer and investor demand) and reputation risk. Both risk types trace back to the same variable, greenhouse gas (GHG) emissions: the more emissions rise, the more physical risk compounds; the harder policy pushes to cut them, the more transition risk compounds.

Risk typeDriverTypical time horizonPrimary data source
Acute physicalExtreme weather eventShort-term (0-5 years)Copernicus Climate Data Store, EEA hazard data
Chronic physicalGradual climate shiftLong-term (15+ years)IPCC AR6 projections, Copernicus
Policy and legal transitionRegulation, carbon pricingMedium-term (5-15 years)EU ETS auction data, CBAM scope
Technology transitionAsset or process obsolescenceMedium-termSector technology roadmaps
Market transitionDemand and investor shiftVariableMarket and customer research
Reputation transitionStakeholder perceptionVariableRating agency and media monitoring

How the two risk types interact

Physical and transition risk move in something close to an inverse relationship over the long run. In a low-mitigation pathway, physical risk keeps compounding because emissions keep rising, but transition risk stays comparatively contained because policy pressure stays weak. In a fast-mitigation pathway, transition risk rises sharply, carbon prices, stranded assets, retooling costs, but physical risk is capped because warming is contained earlier. No credible scenario minimizes both at once; every pathway trades one against the other.

Example: a cement producer that keeps emissions-heavy kilns running the longest carries lower near-term transition risk, no retrofit spend, no carbon-price exposure yet, but higher chronic physical risk exposure across its asset base later in the century. A producer that retrofits early inverts the trade: higher transition cost now, lower physical exposure later. ESRS E1 requires this joint view, not two separate risk lists.

What ESRS E1 requires for each risk type

Under the CSRD, ESRS E1 requires double materiality for both risk types: how the company affects the climate (impact materiality) and how the climate affects the company (financial materiality). E1-9 covers anticipated financial effects from material physical and transition risks and from climate-related opportunities in the same disclosure point, adaptation options for physical risk sit alongside new markets or products for transition risk.

A rough map from the four TCFD transition-risk categories onto ESRS E1:

TCFD categoryESRS E1 disclosure point
Policy and legalE1-2 policies, E1-3 actions and resources
TechnologyE1-1 transition plan, E1-4 targets
MarketE1-9 anticipated financial effects
ReputationSBM-3 material impacts, risks and opportunities

Since Directive (EU) 2026/470 ("Omnibus I") took effect on 18 March 2026, the CSRD reporting obligation applies only to companies exceeding both 1,000 employees and €450 million net turnover, for financial years starting 1 January 2027, first reports due in 2028. Listed SMEs are fully exempt. The CSRD is a directive, so the binding date comes from national law: Germany's implementing act had not been passed as of October 2026. The physical-vs-transition assessment itself is unchanged by Omnibus, only the population of companies required to publish it has shrunk. For the full disclosure workflow, GOV-3 governance oversight, SBM-3 risk listing, and the E1-1 to E1-9 sequence, see my full ESRS and CSRD disclosure guide.

Below the threshold the duty ends, the questions do not. Banks, large customers and tenders keep asking. In the Voluntary Standard (VS), the VSME adopted as Delegated Regulation (EU) 2026/1560, climate risks sit in the comprehensive module under C4, and C4 is outside the value-chain cap: a reporting customer cannot demand it from a supplier with up to 1,000 employees on CSRD grounds, while banks are only encouraged to limit their requests. For a company that has dropped out of scope, I advise keeping the site-level physical risk screening, because that is the piece the bank asks for.

Choosing scenarios and data sources for each risk type

Physical and transition risk need different scenario families and different data. For physical risk, model at least two IPCC pathways, a lower-emission case (RCP4.5) and a high-emission case (RCP8.5), over a horizon of 30 years or more; my RCP vs SSP scenario decision guide covers which pathway fits which sector. For transition risk, use a policy-scenario family such as the NGFS climate scenarios, which model orderly, disorderly and hot-house-world policy paths.

Free public data sources, by risk type:

  • Physical: the Copernicus Climate Data Store for climate projections, the European Environment Agency for hazard maps, IPCC AR6 for pathway assumptions.
  • Transition: EU ETS auction data for carbon-price paths, the CBAM product scope for exposure, NGFS scenario output for macro assumptions.
  • Both: document the model version, spatial or sector resolution, and retrieval date for every source, an auditor needs to reproduce the number, not just read it.
  • Once for several uses: a physical screening built at site level and per hazard serves the EU Taxonomy adaptation screening, ESRS E1-9 and IFRS S2 without a second round of data collection.

Whichever pathway you pick, document why: a single-scenario analysis, one RCP, one policy path, meets the letter of ESRS E1 but hides the spread between best and worst case, usually where the decision-relevant information sits.

Putting a number on each risk

Physical risk is quantified through exposure, hazard and vulnerability: identify the asset value at risk, estimate the probability of a hazard event from historical data and climate projections, then multiply. The worked example under prioritising sites shows the calculation with figures.

Transition risk is quantified through exposure to a policy or market variable: multiply a physical quantity (tonnes of CO₂e, square metres of retrofit-eligible floor space) by a projected price or cost. Example: a company emitting 100,000 tonnes of CO₂e under an assumed carbon price of €100 per tonne faces €10 million in additional annual cost, before any pass-through from suppliers facing the same price.

My position: the harder number to get right is not the physical damage estimate, it is the transition-risk price path, because it depends on a political decision rather than a measured hazard. Most companies I see treat both risk types with the same false precision, one number, no range, when the transition-risk number in particular deserves a wide one.

Both numbers need a documented method, not just a figure: which data source, which scenario, which assumptions, so a reader or an auditor can reproduce the result rather than take it on trust. That documentation, not the number itself, is what ESRS E1 actually audits. One operational note worth keeping: update both scenario sets annually, ideally by October, so there is time to fold new assumptions into a report due the following spring.

How to prioritise physical climate risks across sites

Most site rankings I am handed sort locations by hazard score: red sites first, green sites last. That answers where the climate is worst, not where money spent on adaptation does the most good. I rank adaptation measures by the loss they avoid, net of what they cost. A site with a high hazard score and little worth protecting drops down the list; a mid-risk site where a cheap measure removes most of the expected loss moves to the top.

The sequence, per site and per hazard:

  1. Expected annual loss. Annual probability of a damaging event under each scenario, times the value exposed at the site, times the share of that value an event destroys.
  2. Effect of each measure. Some measures cut the damage per event, such as a hail net over vehicles parked outdoors. Others cut how often the event reaches the asset, such as a backflow valve in front of an underground car park. Treating both as a flat discount on the loss is the most common modelling error I see.
  3. Adaptation return. Avoided annual loss, discounted over the life of the measure, minus its cost. Then one ranking across all sites, not one list per site.
  4. Scenario check. The same ranking under a moderate and a high-emissions pathway. A measure that only pays off under the high pathway is a decision about risk appetite, and the analysis should say so.

A worked example with two measures. The figures are illustrative, not client data; 5% discount rate, 15-year life, maintenance ignored.

Hail net, outdoor vehicle stockBackflow valve, underground car park
What the measure changesDamage per eventFrequency of the event
Annual probability of a damaging event4%2%, with valve 0.4%
Damage per event15% of €2.0 m, with net 3%€150,000
Expected annual loss, without / with€12,000 / €2,400€3,000 / €600
Avoided loss per year€9,600€2,400
Present value of avoided loss€99,600€24,900
Cost of the measure€60,000€4,000
Adaptation return, net€39,600€20,900
Avoided loss per euro invested1.76.2

The net figure puts the hail net first, the ratio puts the valve first. With a fixed adaptation budget the ratio decides, without one the net figure does. That choice belongs in the method note, because it changes the order of the list.

The weakest input is rarely the climate data. Damage functions exist for many asset classes; for floods, the European Commission's Joint Research Centre publishes depth-damage curves by asset type. How much a specific measure reduces that damage is documented far less often. Where I cannot source the effect of a measure, it stays off the ranking and is listed as a data gap, rather than ranked on a default nobody could defend in an audit.

Two more settings decide the result. Physical risk escalates over decades, so a payback window of five years mostly measures today's climate; I calculate to 2050, to 2070 for long-lived buildings, and let the life of the measure cap the return. And avoided damage is one of three returns: a measure can also cut energy cost and emissions, as shading does by lowering cooling demand, which links it to the transition side and the carbon price path above. Side effects such as better working conditions in a heat wave I list without a euro figure, because a monetised co-benefit is usually an invented number.

On a multi-site industrial group in Central Europe, a group-level risk narrative was sent back, while the same substance as a per-site table went through without further questions. Nothing in the analysis had changed; the granularity was the deliverable. For a portfolio of 18 sites, a spreadsheet does this job. Every climate risk analysis I deliver ends in this ranking of adaptation measures by return; whether resilience then actually improves is measured with the climate adaptation KPIs.

FAQs

How should businesses prioritise physical climate risks across locations?

By the loss an adaptation measure avoids, net of its cost, not by hazard score alone. Calculate the expected annual loss per site and hazard, the effect of each measure on damage or frequency, and rank all measures across all sites by discounted return under at least two scenarios. A measure whose effect cannot be sourced belongs on the list of data gaps, not in the ranking.

Is "TCFD physical risk" the same as ESRS E1 physical risk?

Yes, same underlying category. TCFD introduced the physical-risk and transition-risk split; ESRS E1 uses the same split but nests it inside double materiality and the E1-1 to E1-9 disclosure points. A TCFD-aligned risk register maps directly onto an ESRS E1 report.

How does Omnibus I change who has to report on these risks?

Since Directive (EU) 2026/470, CSRD reporting applies only to companies exceeding both 1,000 employees and €450 million net turnover, for financial years starting 1 January 2027, first reports due in 2028. Listed SMEs are exempt. The risk-assessment substance under ESRS E1 is unchanged for those still in scope; below the threshold, banks and large customers keep asking for a site-level physical risk screening.

What are the two main types of physical climate risk?

Acute and chronic. Acute risks are event-driven: floods, storms, heat waves, wildfires. Chronic risks are slow shifts: higher average temperatures, sea-level rise, changing rainfall and growing water stress. TCFD and ESRS E1 use the same split, and both expect it to be assessed site by site, because one company can face river flooding at one plant and water stress at another.

What is an example of a transition risk?

A carbon price that rises faster than planned. A cement or steel producer that budgeted for EU ETS allowances at today's price faces higher operating costs as the cap tightens; that is policy risk, one of the four TCFD transition-risk categories. The other three, with one example each: technology (a gas boiler fleet that loses value once heat pumps become standard), market (a key customer switching to lower-carbon suppliers) and reputation (a lost tender because the buyer asks for a credible transition plan).

Johannes Fiegenbaum

Johannes Fiegenbaum

ESG and sustainability consultant based in Hamburg, specialised in VSME reporting, carbon accounting and climate risk analysis. Works with mid-sized companies, investors and financial institutions, and developed the VSME reporting software VSEasy together with clients.

More about