What Is Climate Risk Modeling?
Climate risk modeling converts climate projections and historical observations into estimates of financial exposure for physical assets, portfolios, and supply chains. Where climate models simulate future atmospheric conditions (temperature, precipitation, sea level), climate risk models take those outputs and ask: what does this mean for a specific building, loan book, or investment portfolio?
The discipline sits at the intersection of climate science, geospatial analysis, and financial risk management. A flood projection alone tells you that water depths may reach 1.2 meters at a given return period. A climate risk model tells you that 1.2 meters of flooding at a warehouse in Houston translates to $4.3 million in expected losses and 12 weeks of business interruption.
Climate risk modeling gained urgency after the Task Force on Climate-related Financial Disclosures (TCFD) published its recommendations in 2017. Regulators, investors, and rating agencies now expect organizations to quantify climate exposure under multiple forward-looking scenarios rather than relying on historical loss data alone.
Physical Risk vs Transition Risk Modeling
Climate risk splits into two categories that require fundamentally different modeling approaches.
Physical Risk Modeling
Physical risk modeling estimates damage from climate-driven hazards at specific locations. These hazards fall into two groups:
Acute hazards are event-driven: riverine and coastal flooding, tropical cyclones, wildfires, and extreme heat events. Models estimate the probability, intensity, and spatial extent of each hazard under current and future climate conditions. A flood model, for example, combines precipitation projections with terrain data, drainage capacity, and land cover to estimate inundation depths at specific coordinates.
Chronic hazards develop gradually: sea level rise, average temperature increase, shifting precipitation patterns, and long-term drought. These affect asset valuations, insurance costs, and operational viability over decades. A coastal property that is viable today may become uninsurable by 2050 under a high-emissions scenario.
Physical risk models typically operate at the asset level. Conducting a rigorous physical climate risk analysis at this level requires precise location data (latitude/longitude), hazard intensity layers at various return periods, and vulnerability functions that translate hazard intensity into damage ratios. The output is usually expressed as expected annual loss (EAL) or climate value at risk (CVaR) under specified SSP scenarios.
Transition Risk Modeling
Transition risk modeling estimates financial exposure from the shift toward a low-carbon economy. Three forces drive transition risk:
Policy and regulation. Carbon pricing, emissions caps, and mandatory disclosure requirements change the cost structure for carbon-intensive businesses. The EU Emissions Trading System, for instance, has pushed carbon permit prices above EUR 60 per tonne, directly affecting energy producers, steel manufacturers, and airlines.
Market shifts. Consumer preferences, investor sentiment, and energy demand patterns are moving away from fossil fuels. Assets tied to coal, oil, and gas face potential devaluation (stranded assets), while clean energy investments attract capital.
Technology disruption. Falling costs for solar, wind, battery storage, and electric vehicles are reshaping competitive dynamics across sectors. Companies that fail to adapt face margin compression and market share loss.
Unlike physical risk models that work at the asset level, transition risk models typically operate at the sector, company, or portfolio level. They combine economic scenario models (such as the Network for Greening the Financial System scenarios) with company-level carbon exposure data to estimate revenue impacts, cost increases, and asset impairment.
Data Sources for Climate Risk Models
Climate risk modeling depends on several data layers, each serving a different function in the modeling chain.
Global climate model outputs. General circulation models (GCMs) from CMIP6 provide the foundational climate projections. These models simulate temperature, precipitation, wind, and sea level under different emissions pathways. The NASA NEX-GDDP-CMIP6 dataset downscales GCM outputs to approximately 25 km resolution, making them practical for regional risk assessment.
Hazard-specific datasets. Flood models use digital elevation models, river gauge data, and land cover maps. Cyclone models use historical track data and sea surface temperatures. Wildfire models use vegetation indices, fuel moisture data, and fire weather indices. Each hazard requires its own specialized data pipeline.
Exposure data. Asset locations (geocoded to latitude/longitude), building characteristics (construction type, number of stories, occupancy), and replacement values. The quality of exposure data often determines the accuracy ceiling of the entire model. A building’s flood risk varies dramatically between the ground floor and the third floor.
Vulnerability functions. Also called damage curves, these translate hazard intensity (e.g., 1.5 meters of flood depth) into damage ratios (e.g., 45% of replacement value). Peer-reviewed curves exist for most hazard-building combinations, published by organizations like JRC, FEMA (HAZUS), and USACE.
Financial data. Property valuations, loan-to-value ratios, insurance coverage, and business interruption costs. For transition risk, this extends to carbon intensity metrics, sector classification, and revenue mix.
How Climate Risk Models Work
Most climate risk models follow a four-step process, regardless of the specific hazard or vendor.
Step 1: Geocode and characterize assets. Every asset in the portfolio gets a precise location (latitude/longitude) and a set of attributes: construction type, occupancy class, replacement value, and elevation. Batch geocoding and automated building classification have reduced this step from weeks to hours for large portfolios.
Step 2: Overlay hazard layers. For each asset location, the model extracts hazard intensity values from gridded datasets. A flood model might pull depth values for 10-year, 50-year, 100-year, and 500-year return periods under both current climate and future climate scenarios (SSP2-4.5 and SSP5-8.5 at 2030 and 2050).
Step 3: Apply vulnerability functions. The model matches each asset’s characteristics to the appropriate damage curve and calculates a damage ratio for each hazard intensity level. A concrete commercial building with 1 meter of flooding might sustain 25% damage to contents and 10% structural damage, while a timber residential building at the same depth might reach 60% total damage. To see how these curves work for a specific building type and flood depth, try the free flood damage calculator.
Step 4: Aggregate and report. Individual asset losses are aggregated across the portfolio. Expected annual loss is calculated by integrating damage across all return periods (probability-weighted). Portfolio-level metrics like probable maximum loss quantify the worst-case scenario at a specified confidence level. Results are typically broken down by hazard type, geography, scenario, and time horizon to support TCFD and CDP disclosure requirements.
Climate Risk Modeling for Banks and Financial Institutions
Banks face specific climate risk modeling requirements that go beyond general corporate risk assessment.
Regulatory stress testing. The European Central Bank, Bank of England, and Federal Reserve have all conducted or mandated climate stress tests. These require banks to model the impact of specified climate scenarios on their loan books, covering both physical damage to collateral and transition risk to borrowers’ revenue streams.
Credit risk integration. Climate risk is increasingly factored into credit underwriting. A mortgage on a property in a high flood-risk zone may require additional provisions. A corporate loan to a coal-dependent energy company carries transition risk that traditional credit models miss. Banks need climate risk modeling outputs that plug directly into their existing credit risk frameworks.
Portfolio-level exposure mapping. Banks with thousands of commercial real estate loans or SME exposures need to screen their entire portfolio for climate concentration risk. Continuuiti’s physical climate risk assessment capabilities, for example, can process portfolio-level exposure across 12 hazards and multiple time horizons, producing the asset-level granularity that regulators expect.
Disclosure compliance. TCFD, CDP, and the EU’s Corporate Sustainability Reporting Directive (CSRD) all require quantified climate risk metrics. Banks need models that produce auditable, scenario-specific outputs in the formats these frameworks prescribe.
How to Evaluate Climate Risk Modeling Tools
The market for climate risk modeling tools ranges from open-source academic frameworks to enterprise SaaS platforms charging six-figure annual fees. Six criteria separate useful tools from marketing exercises.
Hazard coverage. Does the tool cover all relevant physical hazards for your geography? A tool that models only flood risk misses wildfire, cyclone, and heat stress exposure. Look for multi-hazard coverage with the ability to assess compound events (e.g., storm surge combined with riverine flooding).
Scenario and time horizon flexibility. Minimum requirement: SSP2-4.5 (middle-of-the-road) and SSP5-8.5 (high emissions) at baseline, 2030, and 2050. Better tools support additional scenarios and custom time horizons. Verify that the underlying data actually comes from CMIP6 projections rather than simple extrapolations of historical trends.
Spatial resolution. A model that operates at 50 km grid resolution cannot distinguish between a property on a floodplain and one on higher ground 2 km away. Asset-level risk assessment requires hazard data at 250 m resolution or finer for most perils.
Transparency. Can you trace the model output back to specific climate datasets, vulnerability functions, and assumptions? Black-box models that produce a risk score without explaining the methodology create audit and governance problems, especially for regulated institutions.
Data input requirements. Some tools require detailed engineering data for each asset (roof type, foundation depth, distance to coast). Others work with basic inputs (address, occupancy type, replacement value). Match the tool’s data requirements to what your organization can realistically provide at scale.
Output format. The outputs should align with your reporting needs. TCFD metrics, expected annual loss figures, portfolio heat maps, and asset-level drill-downs are table stakes. The best tools produce outputs that integrate with existing risk management and reporting systems without manual reformatting.
Get a Climate Risk Assessment
See how climate hazards affect your locations under multiple scenarios.
Frequently Asked Questions
What is climate risk modeling?
Climate risk modeling translates climate science projections into financial exposure estimates for physical assets and portfolios. It combines climate model outputs with asset-level data and vulnerability functions to calculate expected losses under different scenarios and time horizons.
What data do climate risk models use?
Five main data layers: global climate model outputs (CMIP6 projections), hazard-specific datasets (flood maps, cyclone tracks), exposure data (asset locations and building characteristics), vulnerability functions (damage curves from JRC, FEMA), and financial data (property values, insurance coverage, business interruption costs).
How do banks use climate risk models?
Banks use climate risk models for regulatory stress testing mandated by the ECB, Bank of England, and Federal Reserve. They also integrate climate risk into credit underwriting, screen portfolios for concentration risk, and produce quantified metrics for TCFD and CDP disclosure.
What is the difference between physical and transition risk modeling?
Physical risk modeling estimates damage from climate hazards (floods, wildfires, heat stress) at specific locations. Transition risk modeling estimates exposure from the shift to a low-carbon economy (carbon pricing, stranded assets, technology disruption). Physical risk works at the asset level; transition risk works at the sector or portfolio level.
What tools are available for climate risk modeling?
Tools range from open-source frameworks to enterprise platforms. Evaluate them on hazard coverage, scenario flexibility (SSP2-4.5 and SSP5-8.5 minimum), spatial resolution (250 m or finer), transparency (traceable methodology), data requirements, and output format. The best tools produce TCFD-aligned outputs that integrate with existing risk management systems.
