To accurately understand climate risks, baseline historical understanding of hazard is required and what large-scale factors influence this for different geographic regions. Then as the climate continues to change, an understanding of changes to frequency, severity, exposure, and vulnerability are all required for a multitude of different perils. To avoid an underestimation of future physical climate risks. Further challenges include the accurate representation of extreme events, their compounding and cascading effects, and the integration of non-linearities associated with tipping points in the climate system.
In recognition of this challenge climate risk assessments have experienced amplified attention in both the academic and private spheres and a growth in climate risk services aiming at setting standards and frameworks as well as the provision of comprehensive climate impact information for the private sector and financial institutions.
Therefore, providing a platform to foster interactions between scientists, risk modellers and assessors, economists and financial experts is urgently needed. With the goal of facilitating such dialogue, this session aims at providing a platform for actors from academia and the private sector to exchange information on strategies for assessing climate risk.
The session is organised under three main pillars:
-Physical Climate Risks: Trends, Processes and Modelling
-Identifying and Managing Climate Risks
-Quantifying Damages and Impacts from Climate Risks
We encourage submissions on a wide range of topics including innovative climate risk modeling and model evaluation, damage functions, integrated assessment modelling, bias adjustment and downscaling methods, climate emulators, climate hazard indicators and their projections for specific sectors (e.g. food, energy, insurance, real estate, supply chains), impact data collection and categorization.
Climate risk assessments increasingly rely on the use of complex modelling chains that aim to simulate the interactions between climate-induced changes in hazard, vulnerability and exposure, often over large spatial domains. Due to this high level of complexity, evaluating the impact of uncertain input data and assumptions on modelling results, and therefore the overall model “credibility”, remains a very complex process. In this talk, I will advocate for the use of more structured approaches to quantify and attribute uncertainty in climate risk predictions, discuss the technical and cultural barriers to the adoption of these approaches, and provide some examples of how uncertainty and sensitivity insights can help inform the validation, improvement and use of models - both in academic research and the private sector.
How to cite: Pianosi, F.: Uncertainty in climate risk modelling, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-7977, https://doi.org/10.5194/egusphere-egu26-7977, 2026.
Estimates of the risk of climate change on society rely on historical estimates of true weather conditions and future projections from global climate models (GCMs), which are typically bias-corrected and downscaled before use. Future projections of climate impacts are affected by uncertainty in the underlying climate data through multiple pathways, only some of which are regularly accounted for in the literature. We investigate the importance of the choice of gridded historical data product used to fit impact models and bias-correct and downscale GCMs on the spread in projections of climate impacts. This decision is often either ad hoc in econometric climate impact studies or made for reasons orthogonal to a given product's performance for metrics and regions of interest, despite known limitations of any particular gridded product and difficulties in product evaluation in regions most vulnerable to climate damages.
We re-estimate three climate impact models from the literature, relating exposure to daily mean or max temperature to annual GDP per capita growth, mortality, and payroll, using four different reanalysis products. We then project damages for each dose-response function using a novel ensemble of GCM projections that accounts for all sources of climate uncertainty, bias-corrected and downscaled to the same four reanalyses to estimate this “observational” uncertainty, and incorporating all runs from multiple Large Ensembles of GCMs to estimate model uncertainty and internal variability. This Bias-Corrected and Downscaled Massive Ensemble (BCD-ME) allows us to partition uncertainty in damage projections between model, internal, and reanalysis sources.
We find that the choice of gridded historical data product dominates the spread in future projections of GDP per capita growth, mortality, and payroll at a given Global Warming Level for most parts of the globe, particularly in the mid-latitudes. Since in common practice this source of uncertainty is not considered, existing climate risk assessments likely underestimate uncertainty in future damages, underestimating the Social Cost of Carbon and possibly undercounting the possibility of plausible but extreme damages. We thus recommend that users of climate data test the sensitivity of their results to the choice of historical data product and use products that have been evaluated for the metrics and regions of interest whenever possible, and call for more research into constraining uncertainties about past estimates of the climate.
How to cite: Schwarzwald, K., Lenssen, N., Horton, R., Bonanno, A., and Wagner, G.: The choice of historical data product dominates climate uncertainty in projections of climate impacts in a 2-degree world, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11680, https://doi.org/10.5194/egusphere-egu26-11680, 2026.
A growing number of societal actors rely on high-resolution meteorological information to understand a changing landscape of physical hazards. Within this context, accounting for uncertainty is crucial to quantify and manage risks, but can be challenging given the potential for various sources of uncertainty to manifest differently across use-cases. Here, we combine three state-of-the-art downscaled ensembles to characterize how different uncertainties affect projections of several temperature- and precipitation-based risk metrics across the contiguous United States. We focus on long-term trends of aggregate indices as well as the intensity of rare events with 10- to 100-year return periods. By leveraging new downscaled initial condition ensembles, we characterize the role of internal variability at local scales and estimate its importance relative to other sources of uncertainty. Our results demonstrate systematic differences in patterns of uncertainty between average and extreme indices, across recurrence intervals, and between temperature- and precipitation-derived variables. We show that temperature metrics are more sensitive to the choice of radiative forcing scenario and Earth system model, while internal variability is often dominant for precipitation-based metrics. Additionally, we find that the statistical uncertainty from extreme value distribution fitting can often exceed the uncertainties related to Earth system modeling, particularly at recurrence intervals of 50 years or longer. Our results can provide guidance for researchers and practitioners conducting physical hazard risk assessment.
How to cite: Srikrishnan, V., Lafferty, D., Hartke, S., Sriver, R., Newman, A., Gutmann, E., Lehner, F., and Ullrich, P.: Varying sources of uncertainty in risk-relevant hazard projections, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11093, https://doi.org/10.5194/egusphere-egu26-11093, 2026.
We present CYCLONE, a deep learning framework based on Graph Convolutional Networks (GCNs) developed to predict tropical cyclone–induced coastal storm surge in the North Atlantic basin. The model generates a coastal storm surge peak map associated with a TC in less than one second.
CYCLONE was trained using tropical cyclone tracks well represented in ERA5 (Bourdin et al., 2022) and storm surge simulations generated with GSTM for the period 1980–2022. For the North Atlantic basin, this dataset includes a total of 247 tropical cyclones.
The core of CYCLONE relies on an architecture of Graph Convolutional Network layers. Each tropical cyclone is represented as an independent graph instance, with nodes corresponding to coastal stations and edges defining the spatial connectivity of the coastline. The adjacency matrix with N coastal stations is fixed and shared across storms, allowing the model to learn spatially consistent patterns of surge propagation while remaining transferable across events and domains.
Training was performed using 80% of the available tropical cyclones. 170 tropical cyclones were used for training, while the remaining events did not generate significant storm surge and therefore did not contribute to the gradient computation. The remaining 20% of the storms (47 events) were used for validation.
CYCLONE is a tool capable of providing rapid, large-scale hazard assessments of tropical cyclones, especially in countries or with limited or no technical infrastructure. In this context, CYCLONE facilitates damage assessments and improves tropical cyclones response capabilities, which are essential for insurance, risk management and adaptation planning; key active areas of research in the context of climate change.
How to cite: Odériz, I., Losada, I. J., and Muis, S.: CYCLONE: A superfast large-scale coastal storm surge model for Tropical Cyclones , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13339, https://doi.org/10.5194/egusphere-egu26-13339, 2026.
Catastrophe models are used by the insurance industry to assess the risk from mid-latitude winter storms, a major driver of financial losses across Europe. A major component of these models is the stochastic event set, a catalogue of thousands of storms of sufficient spatial coverage and resolution to be used to support robust risk analysis for a (re)insurer’s property or motor portfolios. The stochastic hazard model must provide a realistic and physically consistent representation of the current storm climatology impacting northern and western Europe. Aon’s Impact Forecasting team have developed a stochastic event set by extracting synthetic events from the output of a Global Circulation Model (GCM). This approach has several advantages as the extracted events are physically consistent, being the product of the physics of the GCM, resulting in a robust storm climatology and clustering depiction.
This study presents a comprehensive approach to calibrate and validate a set of downscaled synthetic storms against gust data from meteorological stations. The storms have been extracted from the LArge Ensemble of Regional climaTe modEl Simulations for EUrope (LAERTES-EU) dataset, providing over 12,000 years of synthetic climate data. The extracted event catalogue includes 62,500 possible winter storm events. The original spatial resolution (~27 km) has been downscaled to 3km. Firstly, a gust climatology of the downscaled storms is constructed and compared against a corresponding gust climatology synthesised from the historical observations of meteorological stations across Europe. A quality-controlled selection of weather stations is used to build the historical event set - spanning between 30 and 60 years, depending on the station. The differences between the synthetic gusts and historical gusts are quantified, analysed and used to build correction coefficients applied to calibrate the synthetic events set.
How to cite: Brocklehurst, A., Georgiadis, A., Braun, L., Ehmele, F., Stadelmaier, K., and Pinto, J. G.: Development of a New Stochastic Event Set for European Wind Storms using GCM Output. , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-21420, https://doi.org/10.5194/egusphere-egu26-21420, 2026.
Hurricanes are among the most destructive and costly natural-hazard related disasters. Post-hurricane field surveys provide crucial real-world observations of building damage and are key to better understanding relationships between structural characteristics and hurricane hazard intensity. However, most existing related studies and readily available datasets primarily focus on residential structures, such that a significant gap remains in the study of commercial building vulnerability to hurricanes. To address this limitation, we develop a dataset capturing wind-related damage caused by Hurricane Ian (2022) to commercial buildings. This dataset integrates property records, satellite and street-level imagery, post-event damage assessments, and estimated hurricane wind speeds, which are spatially linked at the individual building level. It covers commercial buildings in Lee County, Florida, one of the most severely impacted area by Hurricane Ian, and includes 344 unique building records.
Using this dataset, we investigate causal relationships between different building features and wind-induced damage, by employing the Double/Debiased Machine Learning (DML) causal inference framework. Results indicate that building shape, number of stories, roof cover material, building material, and roof shape are, in descending order, the most influential factors affecting damage. For example, buildings with an elongated rectangular shape are associated with an average increase of approximately 34 percentage points in the probability of damage. In contrast, low-rise buildings are associated with an average reduction of approximately 25 percentage points in the probability of damage, relative to mid-rise buildings. These findings provide an important foundation for evaluating and improving hurricane wind vulnerability models and, therefore, hurricane catastrophe risk assessments.
How to cite: Atici, A. T., Cremen, G., Vessey, A. F., Ribeiro, R. Q. C. R., and Iacoletti, S.: Investigating the Key Drivers of Hurricane Wind Damage in Commercial Buildings Using Causal Inference, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13883, https://doi.org/10.5194/egusphere-egu26-13883, 2026.
Rapid development of offshore wind farms in the Northwest Pacific – led by China with over 40 GW of installed capacity – has concentrated high-value infrastructure in one of the world’s most tropical cyclone (TC) active basins. However, widely used vendor natural catastrophe models are primarily designed for land-based exposure and do not adequately represent offshore TC hazard.
In this study, we introduce a framework for assessing TC risk for offshore wind farms. Using stochastic TC track sets, we generate hazard footprints representing maximum wind speeds across offshore sites. These footprints are integrated with industry exposure data to estimate potential damage and financial loss distributions. We further evaluate uncertainty in hazard representation through sensitivity analysis using different TC track sets. Finally, we assess the impact of climate change by incorporating projected shifts in TC intensity and frequency under warming scenarios, highlighting how future climate conditions may alter offshore wind risk profiles.
How to cite: Wang, X., Roemer, T., Vollenbroeker, B., Pissulla, D., Morrison, J., and Hanekop, O.: Assessing Tropical Cyclone Risk for Offshore Wind Farms in the Northwest Pacific Basin, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1539, https://doi.org/10.5194/egusphere-egu26-1539, 2026.
Severe convective storms (SCS), including hail, tornadoes, straight-line winds, lightning, and heavy precipitation, represent a significant and evolving source of climate risk. SCS perils pose significant challenges for sectors such as insurance and finance, where accurate risk quantification is essential for underwriting, portfolio management, and resilience planning. Assessing the risk of these perils requires robust frameworks capable of capturing non-linear dynamics, spatial heterogeneity, and compounding effects. However, current modeling approaches often exhibit limited skills when restricted to narrow hazard scopes (e.g., hail-only) or coarse annual scales, limiting their ability to resolve seasonal and intra-seasonal variability. This research introduces a risk assessment framework that leverages deep learning architectures, specifically, a U-Net model augmented with attention mechanisms, to predict the frequency and severity of SCS perils. The model is trained on high-dimensional interpretable meteorological predictors calculated in-house from reanalysis and climate model data, and georeferenced hazard observations from diverse sources. Attention layers within the U-Net architecture enhance feature localization and interpretability, addressing challenges in modeling rare and spatially complex events critical for risk assessment. The framework produces peril-specific daily probabilities and climatological maps, allowing for modeling cross-peril correlation as well as multi-day outbreaks. By integrating physical understanding with data-driven modeling, this approach offers a scalable and interpretable solution for climate risk assessment to support applications such as underwriting, accumulation management, and risk mitigation.
How to cite: Masello, L. and Panosetti, D.: Advancing Climate Risk Modeling of Severe Convective Storms Through Deep Learning, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1528, https://doi.org/10.5194/egusphere-egu26-1528, 2026.
Addressing climate change loss and damage is a crucial ambition within international climate policy. Given the disproportionate impact of climate change on vulnerable communities, there is a need to develop quantitative tools to support just and equitable decisions surrounding financing and redress for loss and damage. However, the complexity of climate change impacts and the challenging academic and political discourse surrounding loss and damage mean a standardised quantitative framework has not been established.
Here we discuss how catastrophe risk models can be used as flexible quantitative tools to help address this critical gap in climate policy. We explore their potential to quantify both economic and non-economic losses, and their ability to adapt to integrate key features such as social vulnerability, thus responding to the complex loss and damage space. We illustrate this by exploring the change in inland flood risk under climate change for three Global South case study regions: Chikwawa in Malawi, Hanoi in Vietnam, and Cagayan in the Philippines. We estimate the risk to three exposure types with both economic and non-economic implications: residential buildings, agricultural crops, and population. Overall, our results show that catastrophe models can produce meaningful, context-specific insights into climate change loss and damage that can guide decisions surrounding adaptation and financing, while highlighting substantial scope for further development across exposure types, risk metrics, and climate change scenarios.
We also highlight some of the key questions revealed during this research and propose directions for future applications of catastrophe models in the loss and damage space, whilst acknowledging important limitations and climate model uncertainties that should be integrated in future work. Finally, we argue that collaboration across sectors – including academia, industry, and local communities – is fundamental to using catastrophe models to contribute appropriately and justly to addressing loss and damage.
How to cite: Galloway, E., Massam, A., Allard, J., Oldham, P., Sarailidis, G., Catto, J., Germond-Duret, C., and Young, P.: Catastrophe risk models as quantitative tools for climate change loss and damage, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-16994, https://doi.org/10.5194/egusphere-egu26-16994, 2026.
Tropical cyclones pose serious threats to human society and ecosystems. Freely available tropical cyclone are typically calibrated using country-level impacts from EM-DAT, which limits their applications for local-scale risk assessment.
Here we present a new, sub-national set of tropical cyclone damage functions based on an unprecedented tropical cyclone damage dataset. First, we develop Wikimpacts 2.0, an expanded version of the publicly available Wikimpacts 1.0 database. The updated database incorporates non-English Wikipedia articles, multi-event articles, and tables and lists from English Wikipedia. After removing duplicates, Wikimpacts 2.0 contains 7,538 events for seven hazard types (Extratropical Storm/Cyclone, Tropical Storm/Cyclone, Extreme Temperature, Wildfire, Flood, Tornado and Drought) , compared with 2,928 in Wikimpacts 1.0. For tropical cyclones, our new dataset represents the largest collection of publicly available damage information.
Second, we re-calibrate tropical cyclone damage functions from Eberenz et al 2021 using 1,114 events with sub-national impact data over 2000–2024 from Wikimpacts 2.0. For damage-function calibration, we first match Wikimpacts events to IBTrACS records, yielding 1,114 matched events out of 1,869 IBTrACS tropical cyclones with landfall. We then compute annual exposure layers for 2000–2024 using the LitPop module in CLIMADA, generating one exposure layer per year for the calibration process. We calibrate damage functions at two spatial scales. At the national level, we use country-level impacts; for each country affected by an event, we compute a damage function. At the sub-national level, we aggregate impacts to administrative level 1 units (states/provinces) and compute a damage function for each unit. Thus, each event yields a set of damage functions across affected regions. These functions will enable improved local-scale risk assessments.
How to cite: Li, N., M. Kropf, C., Riedel, L., N. Bresch, D., Quilcaille, Y., Zahra, S., Madruga de Brito, M., Worou, K., Jezequel, A., Kurfali, M., Nivre, J., Zscheischler, J., Messori, G., and Thiery, W.: A new set of tropical cyclone damage functions calibrated with the Wikimpacts 2.0 database and CLIMADA ensemble-of-strategies method, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-11328, https://doi.org/10.5194/egusphere-egu26-11328, 2026.
Winter windstorms rank as one of Europe's deadliest and most damaging natural disasters. To model the impacts of these windstorms, surface wind data can be incorporated into climate risk models to derive estimates of natural hazard-related impacts on natural or socio-economic systems. In CLIMADA, risk from a natural hazard can be modelled as the convolution between three components - hazard, exposure, and vulnerability. The vulnerability component links the hazard and exposure components to give total impact that can be approximated through functional relationships called vulnerability curves (or impact functions in CLIMADA). Advancing the science of impact estimation from windstorms is imperative for mitigation and management of changing climate risks, and this relies on appropriate calibration of the vulnerability curve. To this end, in this study, we calibrate a popular impact function from Schwierz et al. (2010) using impact data from two types of sources: open-source (EM-DAT/XWS) and proprietary (PERILS). Results indicate substantial differences between the calibrated vulnerability curves and highlight the importance of the type of recorded disaster data used in calibration. Furthermore, for each of the aforementioned calibration cases, we discuss the uncertainties associated with the use of different cost functions and optimization techniques in the calibration process. The study brings forth how data and method choices influence vulnerability curves, helping better understand modelling uncertainty and support the development of more reliable tools for climate risk assessment and adaptation.
How to cite: Mishra, A. N., Messori, G., Riedel, L., Satheesh, A. R., and Pinto, J.: Improving Europe-wide windstorm damage model using insurance loss data, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5915, https://doi.org/10.5194/egusphere-egu26-5915, 2026.
Hurricanes are among the most destructive natural hazards worldwide, posing significant risks to communities and economies. The Saffir–Simpson hurricane wind scale is widely used to communicate hurricane magnitude, but it relies solely on wind speed and has limited predictive skill of potential damages. In this presentation and in a recent paper, we introduce a novel statistical modelling approach that integrates publicly available hazard, exposure, and vulnerability data to more skilfully predict the financial impact of impending landfalling North Atlantic hurricanes.
By applying optimal weights to hurricane hazard, exposure, and vulnerability attributes, our model significantly improves damage predictions, reducing root mean squared error from over $35 billion USD when using the Saffir–Simpson hurricane wind scale to just $7 billion USD when using our new model. This new simple model greatly outperforms conventional single-parameter damage estimates e.g., hurricane Vmax and central pressure (Cp). We also propose a new ' Predictive Hurricane Damage Scale' that indicates Hurricane magnitude as a function of damage. This new scale facilitates clearer communication for financial industries of potential damages from an impending hurricane, whilst being open source. This framework not only enhances understanding of past hurricane impacts but can also help policymakers and stakeholders prepare more effectively in the days preceding a hurricane landfall. The approach underscores the importance of open-source exposure and vulnerability data, which is a necessity for quantifying risk.
How to cite: Vessey, A., Baker, A., Marcellin-Honore, V., and Michelin, J.: Enhancing North Atlantic Hurricane Damage Prediction Through Integration of Hazard, Exposure, and Vulnerability Data, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12432, https://doi.org/10.5194/egusphere-egu26-12432, 2026.
Climate change adaptation requires actionable information at scales relevant to decision-making. We present the development of a climate data platform (https://ccrab.rcec.sinica.edu.tw/) that integrates downscaled climate projections to deliver accessible climate services for diverse users in Taiwan. The platform architecture employs advanced downscaling techniques to transform global climate model outputs into high-resolution datasets, coupled with user-friendly visualization and data access tools that bridge the gap between climate science and practical application. Beyond research applications, the platform addresses growing demand for climate risk data in financial sectors, providing standardized projections that support Task Force on Climate-related Financial Disclosures (TCFD) reporting requirements and climate risk assessments for businesses and financial institutions.
A critical challenge in developing effective climate services lies in meaningful stakeholder engagement. Understanding the diverse needs of decision-makers across sectors, from water resource management to agricultural planning and disaster risk reduction, requires sustained dialogue and iterative co-design processes. This engagement is complicated by the technical complexity of climate data, varying levels of climate literacy among users, and the need to balance scientific rigor with practical usability.
Determining optimal spatiotemporal resolution presents a fundamental technical and practical challenge, particularly acute in regions with steep topographic features such as Taiwan, Japan, and the European Alps. In these mountainous terrains, climate variables can vary dramatically over short distances due to elevation gradients, orographic effects, and valley-plain transitions. While stakeholders often request the finest possible resolution to capture these local variations, computational constraints, data storage limitations, and uncertainties inherent in downscaling methods necessitate careful trade-offs. The challenge intensifies when complex topography creates microclimates that even high-resolution models struggle to represent accurately, which is a critical issue for Taiwan, a small island country with rough terrains. We discuss our approach to identifying appropriate resolutions for different applications and regions, considering both scientific validity and stakeholder requirements, while acknowledging that higher resolution igher accuracy in topographically complex areas. The platform ultimately aims to provide climate information that is both credible and usable for adaptation planning and climate risk assessment.
How to cite: Lee, S.-Y., Hsu, H.-H., and Kuo, S.-Y.: Building a Climate Data Platform: Balancing Downscaling Resolution, Stakeholder Needs, and Service Delivery, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-15750, https://doi.org/10.5194/egusphere-egu26-15750, 2026.
Quantification of risk must deal not only with long time-averages but with temporal volatility and recognition of any underlying temporal trends, both of which are often dominated by rare but exceptionally destructive events.
This study will present the results of multiple 100-year simulations of synthetic Atlantic tropical cyclones, forced using output from a global climate model. The generated stochastic tropical cyclone tracks have been converted into insurance losses using a hurricane windfield model and a realistic exposure dataset that returns a reasonable average annual loss for Atlantic hurricane risk.
The work presented will address two topics: firstly, the volatility of results between the 100-year simulations and, secondly, any implication of temporal trends from the same datasets. Both topics will consider the volatility between simulations through the lens of the lifecycle of tropical cyclones in each season: from basin and landfalling storm frequency through to the aggregated seasonal insurance losses to identify the points along the lifecycle of storms where most volatility arises.
How to cite: Dixon, R. and Emanuel, K.: Volatility in Tropical Cyclone Losses, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-14582, https://doi.org/10.5194/egusphere-egu26-14582, 2026.
As global temperatures rise, extreme weather events are increasingly causing damages across human health, infrastructure, agriculture, and the broader economy. The science of event attribution is evolving to include estimates of economic damages attributable to climate change in addition to physical impacts. A key challenge in this field is to create physically consistent and high-resolution counterfactuals which can be used to estimate to attributable losses.
Here, we analyse the precipitation-driven impacts of Storm Irene in August 2011 when it was undergoing extratropical transition in the Northeastern United States. Across the Northeast United States, this storm caused rainfall of up to 180 mm within a few hours, leading to fluvial and pluvial flooding with catastrophic consequences that caused more than $1.3 billion in property damages in the state of Vermont alone.
Our method enables linking economic damages attributable to climate change to meteorological drivers through a direct modelling chain by combining an operational weather forecasting model, hydrodynamic model, and economic damage model.
This research underscores the potential of interdisciplinary attribution methodologies to inform climate risk assessments in insurance and provide an evidentiary basis for climate-related liability.
How to cite: Ermis, S., Ginesta, M., Wetzer, T., Franta, B., and Stuart-Smith, R.: Economic damages attributable to climate change in the Northeastern United States from 2011 Storm Irene, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-4072, https://doi.org/10.5194/egusphere-egu26-4072, 2026.
Integrated assessment models have long been used for systemic energy policy design and assessment, but they remain limited when incorporating climate impact feedback typically resorting to discrete SSP-RCP combinations with limited flexibility to evaluate different emission trajectories. Where climate impacts are incorporated, they typically use sector-specific ad-hoc methods, making it difficult to distinguish substantive differences across impact channels from artifacts of implementation. This is especially important as the compound effects of climate impacts and their cascading consequences become more salient. Here we bring forward a standardized abstraction for flexible climate impact emulation which allows for easy extension suitable for a general class of integrated assessment models and climate impact drivers. Our novel contribution is via the use of the Rapid Impact Model Emulator (RIME) which allows the emulation of climate impacts based on global warming levels. In conjunction with simple climate model MAGICC we can emulate impacts for two climate impact channels: reductions in usable thermoelectric power plant capacity due to rising temperature and buildings energy demand changes via reduced heating demand and increased cooling demand under warming. These reflect supply and demand side climate impacts. Emulation spans emission projections from a granular range of full-century carbon budgets, reflecting the diversity in mitigation scenario outcomes and allows for quantifications of small temperature differences in system costs. In isolation, the reductions in thermoelectric plant capacity due to changes in hydroclimatic conditions cause a 20% reduction in freshwater-based cooling technologies as well as a global 2% reduction in coal energy between 1.7C and 2.7C warming scenarios.
However, the joint impact of both drivers influences the technological choices with increased adoption of renewable energy sources with 15 EJ less coal capacity than under the effect of increased energy demand alone, between the same warming levels. This is a consequence of cooling constraints limiting the scalability of thermoelectric powerplants in years where buildings energy demand rises most. The first-best model response then takes account of infrastructure lock-ins engendered and drives the overall energy system into a different path with less thermoelectric power generation across the time horizon. This demonstrates the potential and importance of considering climate impact drivers as well as establishing the viability of flexible impact emulation in Integrated Assessment Models.
How to cite: Raghunathan, V., Vinca, A., Byers, E., and Krey, V.: Flexible climate impact emulation of thermoelectric power plant cooling constraints and buildings energy demand in integrated assessment modelling. , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20453, https://doi.org/10.5194/egusphere-egu26-20453, 2026.
Extreme weather events such as hurricanes exert increasing pressure on communities in hazard-prone areas and on the systems designed to protect them. Insurance serves as a primary risk-transfer mechanism, providing financial security for homeowners and supporting community resilience. Yet, behind this first layer of protection lies a complex web of reinsurers, capital markets, and public institutions that collectively absorb and redistribute disaster risk. Intensifying climate hazards, continued coastal development, and evolving market dynamics threaten the stability of this network.
In this study, we develop a risk-propagation model to assess whether single or sequential hurricanes striking Florida could generate systemic financial stress across the property insurance system. The model links physics-based, probabilistic simulations of hurricane wind and flood losses with detailed data on the Florida residential insurance market, its backstop mechanisms, and regulatory frameworks. We examine how losses cascade through interconnected entities under the present-day status quo, under future climate conditions, and when accounting for evolving market dynamics and adaptation measures, revealing who ultimately bears the bulk of catastrophe risk.
How to cite: Meiler, S.: Who bears the risk? Stress-testing insurance system stability under evolving risks, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1563, https://doi.org/10.5194/egusphere-egu26-1563, 2026.
Humid heat can reduce local labour productivity, dampening production in most economic sectors. These regional production disruptions may propagate through global supply chains, which result in spillover effects and induce macroeconomic losses. In a warming climate, characterised by an increasing frequency and intensity of heatwaves, these spillover risks to global producers and consumers due to humid-heat-induced production disruptions remain unclear. By integrating a recently released wet-bulb globe temperature dataset into the well-established agent-based economic loss-propagation model Acclimate, we assess direct regional production losses as well as resulting indirect losses and risks to different regional sectors within global supply chains under present-day climate and future warming scenarios. We identify key producers and consumers that are particularly prone to supply chain disruptions and highlight the heterogeneity of risks across different income groups within and between countries. These results can support the design of region-specific risk management strategies for humid heat and guide the prioritisation of adaptation investments toward the most vulnerable sectors and regions.
How to cite: Wu, X., Kuhla, K., and Xie, Y.: Global spillover risks from humid-heat-induced production disruptions , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-15237, https://doi.org/10.5194/egusphere-egu26-15237, 2026.
Glacier retreat is reshaping water availability in tropical mountain catchments, with direct consequences for water-dependent economic activities. This study quantifies the economic losses attributable specifically to glacier retreat in the hydropower and irrigation sectors of the Santa River Basin (Peru) for two future horizons (2040–2050 and 2090–2100) under three climate and socioeconomic scenarios (SSP1-2.6, SSP3-7.0, SSP5-8.5).
We combine a lumped hydrological model that explicitly represents glacier melt with an economic assessment of irrigated agriculture and hydropower production. To isolate the effect of glacier retreat from concurrent climate and socioeconomic changes, we apply a three-stage framework: (i) simulation of historical conditions (1981–2020) to calibrate and validate hydrology and define a baseline (2010–2020); (ii) future simulations driven by climate and socioeconomic scenarios with glacier extent fixed at baseline conditions; and (iii) future simulations including scenario-consistent glacier retreat. Economic losses due to glacier retreat are derived from the difference between stages (ii) and (iii). Agriculture losses are estimated from crop-specific water–production relationships for the main crops in the Ancash region, while hydropower losses are assessed for the Cañón del Pato plant based on flow-dependent turbine operation and electricity prices. Environmental flow requirements are included in the study.
Results show that glacier retreat reduces runoff in all months and scenarios, with the strongest impacts during the dry season. By mid-century, glacier retreat alone increases economic losses by ~8% in agriculture and ~15% in hydropower relative to futures without glacier change; by the end of the century these increases reach ~15% and ~30%, respectively. Averaged across scenarios, glacier retreat generates additional losses of about USD 170 million by 2050 and USD 360 million by 2100. Losses are highly scenario-dependent: under SSP5-8.5, mid-century losses are comparable to late-century losses under SSP1-2.6, highlighting the accelerating economic costs of high-emission pathways.
Our findings demonstrate that glacier retreat is not a marginal hydrological signal but a major economic driver in glacier-fed basins, with implications for long-term water and development planning.
How to cite: Muñoz, R., Drenkhan, F., and Huggel, C.: The price of glacier retreat in the water resources sector, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13654, https://doi.org/10.5194/egusphere-egu26-13654, 2026.
The usefulness of predictions of physical climate risks to the financial sector is now appreciated but climate forecasting can also learn from the ability of financial markets to aggregate distributed information and expertise.
CRUCIAL is an initiative that uses “prediction markets” — markets designed to discover and synthesize information rather than transfer assets or risks — to elicit and aggregate expert judgements about climate-related risks. Teams of expert participants, from academia and the private sector, are allocated credits which they can use to trade contracts tied to climate-related outcomes. The prices of these contracts can be interpreted as probabilities that evolve in real time as new information becomes available to participants.
Using prediction markets to aggregate climate forecasts means that the users of the forecasts do not have to select a single provider. This is an important feature because, for longer horizon forecasts, providers cannot demonstrate their competence with a statistically meaningful track record of accurate predictions. Instead, prediction markets directly reward forecasters for the contributions they make to improving the accuracy of collective forecasts.
CRUCIAL’s platform has been used to run markets that predict seasonal temperatures and rainfall, crop yields, El Niño-Southern Oscillation and Atlantic hurricane activity for horizons of up to 18 months ahead. These pilot markets produced forecasts that were consistent with good probabilistic calibration (reliability). CRUCIAL plans further markets with longer prediction horizons.
In a world where historic statistics of climate risks are not necessarily a good indication of future risks, prediction markets provide a mechanism which can combine information from historical data, climate models, and more tacit forms of expertise into quantitative probabilistic forecasts. Prediction markets have the potential to become a new type of scientific institution for synthesizing, summarizing and disseminating diverse climate expertise and different modelling approaches. Prediction markets can also be used to allocate funding for climate forecasting more efficiently than peer-reviewed grants. Such markets could allow experts from many different disciplines and both academia and the private sector to contribute effectively to the generation of probabilistic predictions of physical climate risks.
How to cite: Roulston, M. and Kaivanto, K.: A market mechanism for synthesizing predictions of physical climate risks, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-419, https://doi.org/10.5194/egusphere-egu26-419, 2026.
The expansion of global renewable energy capacity is critical for the net-zero transition, yet traditional top-down risk assessments often obscure the specific physical hazards threatening individual assets. To construct truly resilient portfolios, risk managers and portfolio investors require a bottom-up risk assessment framework that aggregates granular, asset-level exposures into a comprehensive financial view. We applied this bottom-up methodology to a global portfolio of utility-scale wind and solar assets with capacities exceeding 20 MW, from the Global Energy Monitor’s Global Wind and Solar Power Trackers database.
Our methodology moves beyond regional averages to model asset-level risk based on specific geolocation and technology types. For solar photovoltaics, we model future power yield by calculating solar cell temperatures at the module level, derived from ambient temperature, incident shortwave radiation, and wind-driven cooling. This allows for precise estimation of temperature-dependent efficiency losses and thermal degradation. For wind energy, bias-corrected wind projections are extrapolated to turbine-specific hub heights, dynamically adjusting power curves and capacity factors. We further refine this bottom-up analysis by incorporating first-principles damage functions for wind and heat impacts on critical components, calibrated against industry-informed damage thresholds.
Our analysis highlights significant regional disparities: while 2030 yield projections in North America and Europe remain relatively stable (showing negligible median deviations of <0.1%), Asia and South America face severe exposure to heat-induced component damage under RCP 8.5, with projected heat damages exceeding 8% and total climate losses in Asia surpassing 20%. These findings represent a critical step towards integrating physical climate science directly into financial asset management. By granulating risk at the asset level, we are advancing the capability to identify optimal locations for technology upgrades and re-energization strategies that are intrinsically resilient to future climate states. Ultimately, this work advances the shift from static historical baselines to dynamic, forward-looking risk assessments. By quantifying these physical constraints, we support investment strategies that ensure the long-term bankability and systemic resilience of the global renewable energy transition.
How to cite: Ferrer, J. V., Remke, T., Mildenberger, M., and Sánchez, L. A.: Quantifying Physical Climate Risk in Renewable Portfolios: Future Yield, Damage, and Financial Impact, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-21273, https://doi.org/10.5194/egusphere-egu26-21273, 2026.
Regulatory and supervisory stress tests have become a central tool through which climate scenarios are translated into financial risk assessments in the (re)insurance sectors. Yet despite increasing technical sophistication, and in the context of recently updated supervisory expectations such as the Bank of England Prudential Regulation Authority’s supervisory statement (SS5/25) on climate-related risk management, there is growing concern that these practices may not meaningfully improve organisational resilience or decision-making at the board and executive level. Much of the focus remains on the precise quantification of individual hazards, while systemic, compounding, and strategic climate risks remain underexplored. This raises a critical question: are prevailing climate risk frameworks optimising measurement at the expense of genuine resilience?
We argue that prevailing regulatory approaches to climate risk assessment have narrowed how risk is conceptualised and communicated. Physical risk scenarios typically isolate single peril–region combinations, while transition and litigation risks are assessed independently, obscuring the potential for interacting and cascading impacts. Moreover, the technical complexity of probabilistic modelling can limit accessibility for senior decision-makers, hindering effective governance and long-term strategic planning.
We propose a layered climate risk storyline framework that complements existing quantitative models. Rather than relying on fully probabilistic compounding, the approach uses coherent storylines to explore how physical, transition, litigation, exposure, and Earth-system risks may interact and amplify impacts under plausible climate futures. This enables the examination of complex and systemic risk dynamics while remaining transparent and interpretable for senior decision-makers.
We suggest that storyline-based, compounding risk frameworks offer a more effective bridge between climate science, catastrophe modelling, and strategic decision-making, shifting the focus from precise loss estimation toward resilience. Positioned alongside national climate services and national climate scenario products, this approach highlights the need for closer collaboration between academia, climate scientists, and practitioners to develop scenario frameworks capable of supporting more robust climate resilience in regulated financial sectors.
How to cite: Giani, G., Noacco, V., Wardman, J., McIlwaine, J., Taylor, H., Flanagan, S., and Philp, T.: A layered climate risk storyline framework for climate resilience, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-21773, https://doi.org/10.5194/egusphere-egu26-21773, 2026.
Accurate wildfire prediction is becoming increasingly critical as climate change drives warmer and drier conditions worldwide. The complex, non-linear interactions among meteorological factors, fuel characteristics, and landscape structure make wildfire risk a strong candidate for advanced machine learning (ML) approaches that integrate Earth Observation (EO) and climate data. Recent progress on this front has already led to significant improvement on operational systems, such as the ECMWF wildfire forecast, demonstrating clear advantages over traditional, meteorology-only indicators. However, most current ML models are based on single pixel predictions that lack essential spatial context. This limits their ability to capture how static forest connectivity interacts with dynamic fire processes, including spread, intensity, and likelihood of occurrence. To overcome these constraints, we propose a Convolutional Neural Network (CNN) architecture designed to explicitly learn and exploit the additional predictability from these complex spatial relationships. The model fuses multiscale inputs by processing high-resolution landscape variables (e.g., above-ground biomass, land cover, soil moisture, topography) alongside coarse-resolution meteorological fields. To represent the full spectrum of wildfire risk, we experiment with multiple target variables including probability of burn, fire severity, and fire extent. Through these experiments, the CNN is forced to learn connectivity patterns directly from historical wildfire events. The successful implementation of this approach would constitute a major step toward operational, high-resolution, context-aware wildfire risk mapping, strengthening both early-warning capabilities and long-term resilience planning.
How to cite: Cendagorta, D., Civantos, D., Perpinyà, M., Florindo, C., Huertas, C., Teruel, D., Romero, L., Llort, J., and Peña-Iquierdo, J.: Learning Fire Connectivity: A Convolutional Neural Network for assessing wildfire risk, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-21514, https://doi.org/10.5194/egusphere-egu26-21514, 2026.
There is a growing demand in reinsurance for parametric modeling frameworks that are not only fast and computationally efficient, but also capable of incorporating real-world, forward-looking scenarios based on observable and projected risk drivers. In response, this study proposes an integrated, climate-adjusted framework for natural catastrophe (NatCat) pricing that combines Average Annual Loss (AAL), machine learning-based disaster frequency modeling, growth-rate attribution, and reinsurance pricing metrics. Using country-level hazard and exposure data, Random Forest models are employed to jointly estimate disaster frequencies from observed AALs and, conversely, to infer AALs from modeled disaster frequencies, thereby ensuring internal consistency across pricing components. Growth rates are quantified at both aggregate and hazard-specific levels and projected under climate scenarios for 2030, 2050, and 2100. The proposed framework enables a forward-looking assessment of climate-driven risk evolution and supports risk-based pricing decisions with direct practical applicability for insurers, reinsurers, and public risk pools engaged in underwriting, capital management, and climate-resilient risk transfer mechanisms. The contribution of this study lies in the integration of machine learning-based frequency estimation, climate-adjusted growth-rate attribution, and reinsurance pricing within a single, internally consistent NatCat pricing framework, rather than in the development of new hazard or climate models.
How to cite: Necesito, I., Lee, J., Lee, S., Kim, S., and Kim, H. S.: Climate-Adjusted Machine Learning-Driven (Re)Insurance Pricing Using Future Projections and Disaster Frequency-Average Annual Loss Dynamics , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-15959, https://doi.org/10.5194/egusphere-egu26-15959, 2026.
Climate science depends on hierarchies of models to understand and to predict climatic variability across spatiotemporal scales. Similarly, macroeconomics after the 2008 financial crisis increasingly employs a plurality of models with distinct aims. Both disciplines often rest on linearity assumptions to model the evolution of averaged quantities over the long-term. Current Integrated Assessment Models (IAMs), however, rely almost exclusively on the latter models, for both the economic and climatic systems. Yet, nonequilibrium and short-term dynamics shape both the risks of climate change and the strategies for their management in the long-term. Neglected interactive climate–economy phenomena – specifically the volatility of commodity prices – are likewise crucial for the stability and growth of developing countries.
We therefore present a minimal data-driven coupled model of the El Niño-Southern Oscillation and the macroeconomy. Crucially, the non-equilibrium economic model reveals a tradeoff between structural stability and resilience: economic management that dampens the amplitude of endogenous fluctuations increases the economy's sensitivity to exogenous shocks. The coupled model reproduces the multiscale oscillatory variability that is observed in the prices of several tropical commodities. These results demonstrate the importance of IAMs that accurately represent the full spectrum of time scales in both the economic and climatic systems for the effective management and understanding of commodity price variability and, more generally, of climate risks.
How to cite: Ohara, D. and Ghil, M.: Minimal modelling of non-equilibrium dynamics in coupled climate–economy systems, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-15882, https://doi.org/10.5194/egusphere-egu26-15882, 2026.
The increasing frequency and severity of climate-related events pose significant challenges to financial institutions, municipalities, and asset owners. As insurers, it is crucial to deepen our understanding of the impacts of severe catastrophic events on the Canadian landscape within the context of climate change. This presentation introduces the Climate Risk Manager (CRM), a state-of-the-art tool designed to offer granular, asset-level risk assessments and promote economic adaptation strategies, particularly for credit unions and insurers.
In response to OSFI B15 requirements, which mandate Canadian financial institutions to disclose their climate risk exposures, CRM provides a transparent and customizable solution to meet these regulatory demands. By incorporating advanced catastrophic models and simulating 50,000+ years of climate catalogue events, CRM translates climate events into probable economic losses, illustrating potential impacts. This integration of exposure, vulnerability, and event data empowers financial institutions to make informed decisions regarding mortgage approvals, portfolio diversification, and regulatory compliance, effectively managing climate-related risks while adhering to industry standards.
Through authentic case studies, including a demonstration with a credit union, this presentation will showcase CRM’s capabilities in identifying risk levels and optimizing insurance coverage, thereby supporting strategic decision-making for enhanced economic resilience. The CRM platform features tools such as the Exposure Explorer and Hazard Explorer, which facilitate asset portfolio analysis and risk assessment for floods and wildfires. By generating synthetic historical climate data, CRM delivers comprehensive risk assessments and loss metrics, including expected average loss and the variance of expected quantile loss. Its precision in risk evaluation is particularly beneficial in urban areas, despite data limitations in rural geocoding.
Emphasizing transparency, CRM enables users to backtrack and understand specific results and assumptions, empowering stakeholders to make informed strategic decisions that navigate the complexities of climate change impacts effectively. Looking forward, CRM will evolve by integrating projected climate scenarios and additional natural catastrophe perils (e.g., severe convective storms and hurricanes). This adaptability positions CRM as a critical resource for navigating future climate challenges, ensuring that organizations remain resilient in the face of evolving climate change risks.
How to cite: Xu, J.: From Data to Decisions: Enhancing Economic Resilience through Climate Risk Manager, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-199, https://doi.org/10.5194/egusphere-egu26-199, 2026.
Assessing how physical climate hazards affect borrower solvency and portfolio resilience remains a critical challenge for financial institutions. Existing approaches often focus on single hazards analysis or top down macroeconomic frameworks. Here, we present a practical, scalable framework that enables central banks and financial institutions to quantify loan-level exposure to multiple physical hazards, and to translate those exposures into asset-level financial impacts and, ultimately, into portfolio expected loss estimates.
This multi-risk, micro-level modelling framework, developed jointly together with an emerging markets central bank and a european decelopment agency, maps asset locations to four hazards: floods, heat, drought, and wildfire. It combines established natural catastrophe and climate-impact methods with new, tractable procedures to convert hazard intensity into yield, revenue, and profit shocks. These shocks are then propagated through a Merton-type credit risk model to produce loan- and portfolio-level expected loss estimates. The entire workflow is implemented in an R Shiny application, allowing users to build custom multi-year, multi-hazard scenarios, upload portfolio data, and directly analyse impacts across firms, sectors, and regions.
This framework has been initially designed and calibrated for the profile of a single country. However, its modular structure enables straightforward scaling to new datasets, additional hazards, and new regions. We believe this setup can be particularly valuable to stakeholders and financial institutions, especially those in developing economies, to advance physical risk assessment and understanding, as well as future regulatory exercises.
How to cite: Buller, A., Hayne, M., Gallice, B., and Thomä, J.: Integrating Physical Climate Hazards into Credit Risk: A Multi-Risk Modelling Approach, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1140, https://doi.org/10.5194/egusphere-egu26-1140, 2026.
Hail causes billions in annual insured losses worldwide. It damages solar panels, roofs, vehicles, and crops; creating massive repair costs and operational disruptions. Financial institutions, insurers, and real estate investors face significant exposure to hail-driven losses, which affect portfolio valuations, underwriting decisions, and asset protection strategies. This hazard triggers immediate insurance claims, jeopardises infrastructure investments, and disrupts supply chains; making it critical for enterprise risk management. As climate change impacts severe weather patterns, businesses need forward-looking hail risk information and not just historical data.
We present a new hail risk model developed by Climate X, featuring future projections across different shared socioeconomic pathways (SSPs) for the continental United States. Our model integrates baseline hail hazard data with climate projection methodologies to assess risk under multiple future scenarios. The framework combines high-resolution meteorological data with vulnerability curves based on asset-specific characteristics to quantify direct physical damage across infrastructure and commercial, industrial, and residential buildings.
The model provides risk assessment at both asset and portfolio levels across multiple return periods, enabling stakeholders to evaluate present-day exposure and future climate scenarios. By incorporating SSP-based projections, our approach addresses the limitations of historical-only assessments and provides actionable intelligence for climate adaptation planning and risk management strategies in a changing climate.
How to cite: Ng, K., Larson, E., Leach, N., Ramsamy, L., and Starr, A.: Climate-Driven Hail Risk Projections for the Continental United States, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19858, https://doi.org/10.5194/egusphere-egu26-19858, 2026.
Climate change is expected to significantly affect insurers’ liabilities through an increase in claims associated with more frequent and intense meteorological hazards. These include both extreme events such as tropical cyclones and storms, and more recurrent phenomena such as floods and droughts. In this context, the mutual insurance group Covéa conducts climate-impact studies in collaboration with the specialized climate-risk Hydroclimat company.
The study focuses on the projected evolution of coastal flooding risk in France by 2050. According to the HANZE database (Paprotny D., 2024), between 1950 and 2020, approximately one-third of flood events in France involved a coastal flooding component. Major historical events, such as Storm Xynthia in 2010 and the October 1987 storm, resulted in insured losses of €660 million and €1.5 billion (CCR, 2023), highlighting the significant financial exposure associated with coastal hazards.
To anticipate future impacts, Hydroclimat produced coastal flood-extent maps based on CMIP6 climate projections, integrating existing coastal protection systems within a hydro-geomorphological modelling framework. Exposure and vulnerability analyses were conducted using Covéa’s national residential building database. These results provide an initial assessment of the projected increase in the number of exposed residential buildings and the associated insured losses by mid-century.
This work contributes to a better understanding of future coastal flood risk under climate change and supports insurers in adapting risk assessment and portfolio management strategies to evolving coastal hazards.
References
Paprotny, D. (2024) - HANZE catalogue of modelled and historical floods in Europe, 1950–2020 (v1.2) https://doi.org/10.5281/zenodo.12635205
Caisse Centrale de Réassurance (2023) – Risque de submersion sur la côte atlantique : l’analyse CCR – https://www.ccr.fr/submersion-marine-cote-atlantique-scenario-ccr/ [last access : 2025/12/15]
How to cite: Terrier, M., Lambert, A., Troin, M., and Poudret, B.: Projected coastal flood impacts in France by 2050 using CMIP6 climate projections, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-2732, https://doi.org/10.5194/egusphere-egu26-2732, 2026.
Climate change is expected to intensify extreme precipitation, increasing future flood-related losses. Yet, prioritizing adaptation remains challenging without credible estimates of the financial impacts of physical climate risk. This study develops an integrated analytical framework to quantify flood-induced financial losses in Taiwan, specifically focusing on the semiconductor, cement, petrochemical, and steel industries. The framework translates climate-driven hazard changes into asset-level value impacts for these critical industrial facilities.
The methodology integrates historical station observations with statistically downscaled precipitation projections from AR6 GCMs. Future daily rainfall is simulated using a multi-site stochastic weather generator (MultiWG). These series are then disaggregated to hourly rainfall using a feature-vector-based k-nearest neighbors (KNN) resampling approach. While general scenarios rely on GCM simulations, this study augments the stress testing framework with bias-corrected AR5 typhoon dynamic downscaling data to better capture extreme event dynamics at higher spatial resolutions. To bridge the gap between rainfall and flood impacts, ten temporal patterns from Taiwan’s Water Resources Agency (WRA) are utilized to estimate scenario-specific frequencies of extreme rainfall. Inverse distance weighting (IDW) is subsequently applied to interpolate location-specific extreme-rainfall frequencies to estimate localized inundation depths based on WRA flood potential maps. WRA depth–damage curves are then overlaid to estimate expected asset losses over 20-year horizons for a historical baseline (1995–2015) and three future periods (2021–2040, 2041–2060, and 2061–2080) under multiple climate scenarios.
Rather than focusing on absolute financial loss figures, this study emphasizes a comparative analysis of average annual losses and tail-risk impacts, quantified through Value-at-Risk (VaR), across the selected industrial sectors. By mapping these quantified risks onto financial statement line items, the framework supports decision-useful reporting and evaluates system stability under extreme events through climate stress testing. Ultimately, this framework facilitates sensitivity analysis to identify priority adaptation targets and optimize investment portfolios. These outputs strengthen TCFD-aligned disclosure by offering a transparent and defensible basis for communicating physical risk and adaptation actions in the industrial sector.
How to cite: Wu, H.: Quantifying Physical Climate Risks for Key Industrial Sectors in Taiwan: A Financial Impact Assessment of Flood Hazards under Multiple Climate Scenarios, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-3704, https://doi.org/10.5194/egusphere-egu26-3704, 2026.
Climate change has intensified extreme rainfall events, increasing flood risks at the local level. To support evidence-based flood management, this study develops a flood risk model based on a two-stage regression structure. The first stage develops a nonlinear flood damage function using daily maximum rainfall as the independent variable. The second stage employs machine learning to relate the coefficients of the flood damage function to flood mitigation policy options, including retention reservoir ratio, pumping capacity ratio, and river channel improvement ratio. This second-stage function operates as a policy evaluation module, enabling assessment of how policy interventions affect flood damage mitigation. The model was developed for 228 municipalities across South Korea using 24 years of historical flood records from 1998 to 2021. The model offers two key capabilities: estimating economic flood damage from rainfall input and comparing economic damage across different policy options. To assess climate change impacts and mitigation effects of policy options, future rainfall projections from the WRF climate model under SSP2-4.5 and SSP5-8.5 scenarios were applied. The analysis indicates that integrated policy interventions could reduce future economic losses by approximately 34.92% under SSP2-4.5 and 1.62% under SSP5-8.5 compared to baseline scenarios. Model development is expected to be completed by 2026, with a web-based platform scheduled for deployment in 2027–2028. Once operational, the platform will enable local governments to assess flood risks and evaluate policy options tailored to their specific conditions, providing practical decision support for climate-resilient flood management.
Acknowledgement
This work was supported by Korea Environment Industry & Technology Institute(KEITI) through Climate Change R&D Project for New Climate Regime, funded by Korea Ministry of Environment(MOE)(grant number RS-2022-KE002152)
How to cite: Park, H., Jee, H. W., and Seo, S. B.: A Two-Stage Regression Framework for Assessing Municipal Flood Risks and Mitigation Policy Effectiveness under Climate Change , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-4710, https://doi.org/10.5194/egusphere-egu26-4710, 2026.
Hydrological models are essential tools for generating streamflow estimates across various scales. While the choice of model structure is often scrutinized, the spatial resolution at which these models operate is a critical factor that directly influences the accuracy and representation of hydrological processes (Hoch et al., 2023; van Jaarsveld et al., 2025): coarser resolutions may fail to capture localized runoff dynamics, whereas finer scales offer better precision at the cost of computational intensity.
One application of hydrological models is to identify and group discharge peaks into event catalogues. These catalogues are integral components of catastrophe (CAT) models, used by the insurance and disaster-management sectors to quantify their portfolio risk and guide underwriting. However, the spatial resolution of the underlying hydrological model may introduce uncertainty into this process: discrepancies in streamflow timing and magnitude resulting from resolution choices may alter how events are clustered, potentially leading to variations in the frequency and severity of events recorded in an event catalogue.
This study presents a sensitivity analysis evaluating the impact of varying model resolutions of the hydrological model Wflow on both streamflow estimations and the subsequent generation of event catalogues. By comparing model outputs across multiple spatial resolutions in the UK and Ireland, we assess the degree of (dis-)agreement in event identification and clustering. Our results aim to shed light on how spatial discretization choices propagate through the risk-modelling chain, ultimately affecting the reliability of flood impact assessments and financial risk projections.
Hoch, J. M., Sutanudjaja, E. H., Wanders, N., Van Beek, R. L. P. H., and Bierkens, M. F. P.: Hyper-resolution PCR-GLOBWB: opportunities and challenges from refining model spatial resolution to 1 km over the European continent, Hydrol. Earth Syst. Sci., 27, 1383–1401, https://doi.org/10.5194/hess-27-1383-2023, 2023.
van Jaarsveld, B., Wanders, N., Sutanudjaja, E. H., Hoch, J., Droppers, B., Janzing, J., van Beek, R. L. P. H., and Bierkens, M. F. P.: A first attempt to model global hydrology at hyper-resolution, Earth Syst. Dynam., 16, 29–54, https://doi.org/10.5194/esd-16-29-2025, 2025
How to cite: Hoch, J., Buitink, J., Marshall, A., and Addor, N.: The impact of hydrological model resolution on streamflow estimation and catastrophe model event clustering, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-4924, https://doi.org/10.5194/egusphere-egu26-4924, 2026.
The HuT (The Human-Tech Nexus) project, funded by Horizon Europe initiative, is focused on risk assessment and disaster risk reduction for distinct types of hazards (wildfires, landslide, droughts, etc.), over the European territory, by means of a series of demonstrators representing a multi-hazard arena. In the framework of this project, we present an innovative Catastrophe Bond (Cat Bond) designed to enhance disaster risk reduction strategies. Cat Bonds are a key financial instrument for transferring the risk of extreme events from insurers to capital markets, thereby increasing resilience and reducing the economic impact of disasters. Our approach lies in the use of the recently developed E3CI (European Extreme Events Climate Index) as the trigger mechanism for the bond. The E3CI is a suite of indicators designed to monitor and quantify the occurrence and intensity of climate extremes across Europe. It integrates multiple variables into a single, scientifically robust metric, enabling consistent and transparent assessment of climate-related risks. By using E3CI as the trigger for our Cat Bond, we ensure that payouts are based on objective, observed climate conditions rather than loss estimates, improving reliability and fairness in risk transfer mechanisms. The coupon here reckoned for the Cat Bond are based on hypothetical portfolios over the Italian territory. The proposed Cat Bond ensures transparency, objectivity, and a strong link to observed climate extremes. This solution represents an interesting case study in integrating climate science into risk financing solutions, supporting both insurers and communities in managing the growing risks associated with climate change.
How to cite: Lo Conti, F., Gallotti, G., Tirri, A., Santoro, A., Mangieri, A., Rianna, G., and Calvello, M.: From Climate Extremes to Financial Resilience: E3CI-Based Catastrophe Bond, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5197, https://doi.org/10.5194/egusphere-egu26-5197, 2026.
Hazards such as storms can create multiple perils, such as windstorms and floods, that have correlated annual losses. To better understand the drivers of such correlations, this study explores three collective risk frameworks with varying complexity.
Mathematical expressions are derived from the assumption frameworks to explain how this correlation depends on parameters such as event dispersion (clustering), and the joint distribution of the two hazard variables. Hazard variables are first assumed independent, inducing a positive correlation due to the shared positive dependence on the total number of events. The next framework allows for correlation between the hazard variables, which can then capture negative correlation between accumulated losses. The final framework builds on this by allowing for between-year correlation caused by interannual modulation of the hazard variables.
These frameworks are illustrated using European windstorm gust speeds and precipitation reanalyses from 1980– 2000. They are used to diagnose why the correlation between annual wind and precipitation severity indices decreases as thresholds are increased. Only the framework with interannual modulation of the hazard variables quantitatively captures the negative correlations over Europe at high thresholds. We propose that one plausible driver for the modulation is the transit time that storms spend near locations.
As this methodology is flexible and can be applied to different aggregation periods and spatial scales, it is applicable to investigations of relationships between other aggregated hazards.
How to cite: Jones, T., Stephenson, D., and Priestley, M.: Collective risk modelling of multi-peril events: correlation of European windstorm gust and precipitation annual severity, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-5238, https://doi.org/10.5194/egusphere-egu26-5238, 2026.
Global flood catastrophe models underpin decisions in insurance, infrastructure planning, and climate adaptation, yet they integrate multiple uncertain components, including terrain representation, flood defences, and climate-driven hazard changes. While each of these elements is known to influence flood risk estimates, there is limited quantitative evidence on their relative importance in controlling loss outcomes at global and regional scales.
Here we apply global sensitivity analysis to a large-scale flood catastrophe modelling framework to assess how loss estimates respond to key modelling and data choices. We systematically vary terrain data type (ASTER/SRTM-derived DSM versus LiDAR-derived DTM), terrain resolution (30 m and 5 m), flood defence representation (defended and undefended views, legacy and updated defence datasets), and climate-driven event sets (baseline, 2°C, 4°C, and 6°C warming scenarios). The analysis is conducted across multiple geographic contexts, including Canada, South Africa, Slovakia, and Germany, capturing a range of topographic, vegetative, and urban conditions.
In our presentation, we highlight the role of sensitivity analysis in flood catastrophe modelling, with a particular focus on terrain data representation. We examine how the choice of terrain data, specifically the transition from DSM to LiDAR-derived DTM, influences variability in modelled flood losses, and how this sensitivity compares with other key assumptions, including climate warming scenarios and flood defence representation. By considering these interacting sources of uncertainty side by side, we demonstrate the value of a multi-parameter sensitivity framework for understanding and prioritising model development choices in flood risk assessment.
These findings demonstrate the value of sensitivity analysis for prioritising data investment and model development in global flood risk modelling. In particular, they suggest that improvements in terrain data quality can yield disproportionately large benefits for loss estimation, with implications for risk pricing, adaptation planning, and climate resilience assessment.
How to cite: Hinks, O., Oldham, P., Eddounia, F., and Young, P.: Assessing the sensitivity of global flood loss estimates to terrain data, defences, and climate change., EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-12881, https://doi.org/10.5194/egusphere-egu26-12881, 2026.
This study outlines the initial steps toward applying the Physical Climate Risk Assessment Methodology (PCRAM) to quantitatively assess and enhance resilience within the agriculture and tourism sectors, which are highly susceptible to climate change and natural disasters such as hail and other perils. Although many risk assessments and models exist globally as detailed as part of this initial review of climate risk analytics for capital in these sectors at a basic level, there exists very little analysis which integrates the direct effects of climate, engineering and socioeconomic change into the operational and capital expenditure. This gap leads to the prevalent issue of undervaluing climate adaptation in investment decisions.
As part of this preliminary study, various risk assessment methods, software and frameworks, such as CLIMAAX and MYRIAD-EU, are reviewed which have been applied to the agritourism industry - given the large influence through a multitude of hazards - both climate driven and geophysical. For this preliminary framework and review the case of agritourism facilities in Northern Italy is identified as a critical pilot region due to its high-value viticulture and the increasing frequency of extreme hail events which threaten both agricultural yields and tourism infrastructure. This case study demonstrates how climate change directly impacts specialized assets such as wineries and farm-stays necessitating a detailed four-step approach.
The first step identifies key assets such as farm infrastructure, wineries, accommodation and crops, and hazards within the agritourism sector. The second step, a materiality assessment, would link climate hazards to potential impacts on these assets, quantifying the severity of effects like crop damage or revenue loss and classifying them as maintenance, performance, or life-cycle costs. The third step, resilience building, identifies and evaluates both structural (e.g. hail nets, retrofitting structures for wind and earthquake) and non-structural (e.g. modified operational plans) interventions, reassessing their impact on the assets. The final step, economic and financial analysis, would compare the financial performance of the three steps to demonstrate the value of investing in resilience. This shows how an initial investment might lead to more stable revenues and a better allocation of costs over the asset's lifespan. Ultimately, this methodology may be scaled to groups of assets and transferred to other susceptible economic sectors as the research evolves.
How to cite: Maier, A., Daniell, J., Kunz, M., Hinz, S., Khazai, B., Schäfer, A., Girard, T., and Brand, J.: Quantifying Resilience: Applying the Physical Climate Risk Assessment Methodology (PCRAM) to Agritourism, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13023, https://doi.org/10.5194/egusphere-egu26-13023, 2026.
Extreme flood events are becoming more frequent and intense, increasingly challenging the protection of urban areas and the resilience of socio-economic systems. Despite the high exposure of the residential sector and the key role of insurance for risk transfer and financial protection, a large share of buildings in many European countries, including Italy, remains uninsured against natural hazards.
Accurately determining flood insurance premiums for the building stock is a complex task that requires a detailed characterization of flood hazard, building exposure, and vulnerability features. This study presents a methodological framework to support the definition of premium benchmarks, with an application to residential buildings in Northern Italy. High-resolution hazard data are combined with tailored damage modelling tools to assess expected losses, which are subsequently translated into insurance premiums using two alternative redistribution strategies. The first, a targeted approach, assigns losses only to buildings in the inundated areas. The second, a mutuality-based approach, redistributes premiums across a broader spatial domain, including all buildings within the affected municipalities. For each strategy, multiple assumptions regarding loss redistribution are examined to explore their impact on premium calculation, while also considering the typical compensation mechanisms adopted in insurance practice.
Finally, flood premiums are compared with estimates derived for seismic risk in high-hazard zones, highlighting both differences and similarities in insurance mechanisms across these two hazards. The results suggest that integrating flood and seismic risk through multi-risk pooling strategies may contribute to a reduction in insurance premiums.
How to cite: Treglia, G., Barucci, E., Cesari, R., D'Aurizio, L., Scorzini, A. R., Simonelli, T., and Molinari, D.: Estimating Flood Insurance Premiums for the residential sector: evidence from Northern Italy, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-13900, https://doi.org/10.5194/egusphere-egu26-13900, 2026.
Since 2011, Sargassum beaching events have intensified across the western Atlantic, driven by the emergence of the Great Atlantic Sargassum Belt. These recurrent strandings generate environmental, economic, and health impacts in the French Caribbean islands, with growing implications for insurers due to disruptions of coastal activities, damage to infrastructure, and increased claims related to pollution and loss of use. In this context, Covéa conducts impact‑oriented studies to better understand how emerging environmental hazards may affect insured assets.
This study examines the long‑term evolution of Sargassum presence around Guadeloupe using satellite observations from the SAREDA dataset (Descloitres et al., 2021 ; Podlejski et al., 2022), provided by the AERIS/ICARE Data and Services Center. MODIS‑derived Sargassum fractional coverage was analyzed from 2003 to 2025 within a 50 km coastal buffer to identify the onset and magnitude of the post‑2018 regime shift. Results show a clear transition from low‑intensity occurrences before 2018 to increasingly severe and frequent events afterward.
To assess potential exposure of insured properties, a geospatial analysis was performed combining building location data, distance‑to‑shore metrics, and recurrent Sargassum accumulation zones derived from satellite observations. This approach identifies residential areas most likely to be affected by future beaching events and provides a first estimate of the associated insurance‑related risks.
This work contributes to a better understanding of Sargassum dynamics in the French Caribbean and supports insurers in integrating emerging environmental hazards into risk assessment and portfolio management strategies.
References
AERIS/ICARE Data and Services Center (2021) – SAREDA dataset. DOI: https://doi.org/10.12770/8fe1cdcb-f4ea-4c81-8543-50f0b39b4eca - last access : 2026/01/15
Descloitres, J., Minghelli, A., Steinmetz, F., Chevalier, C., Chami, M., & Berline, L. (2021). Revisited Estimation of Moderate Resolution Sargassum Fractional Coverage Using Decametric Satellite Data (S2‑MSI). Remote Sensing, 13, 5106. https://doi.org/10.3390/rs13245106
Podlejski, W., Descloitres, J., Chevalier, C., Minghelli, A., Lett, C., & Berline, L. (2022). Filtering out false Sargassum detections using context features. Frontiers in Marine Science, 9:960939. https://doi.org/10.3389/fmars.2022.960939
How to cite: Bouldoyre, C. and Poutier, D.: Mapping Exposure to Sargassum Beaching Events for Insurance Risk Assessment in the French Caribbean, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-17129, https://doi.org/10.5194/egusphere-egu26-17129, 2026.
Winter windstorms are among the costliest natural hazards in Europe, with average annual insured losses estimated at €1.4 billion. In France, they consistently represent the most damaging peril. Estimating windstorm losses remains challenging because they are dominated by rare extreme events and due to the compounded nature of storm activity.
Windstorm losses are typically estimated using vulnerability curves that relate storm intensity to the probability and magnitude of damage. However, windstorms frequently occur in close temporal succession, forming storm clusters. The impacts of such compound events can accumulate, leading to cumulative losses that exceed those associated with isolated storms. While wind-impact vulnerability curves generally perform well, they do not account for the role of storm clustering in shaping damage occurrence and intensity. Improving the representation of clustered storm impacts could therefore refine risk characterisation, enhance loss estimation for both individual and compound events, and increase flexibility in reinsurance design.
Using the portfolio of Generali France as a case study, we investigate the role of storm clustering in wind-related insurance losses. Losses are first associated with individual storm tracks, and storm clusters are defined as sequences of damaging events separated by less than 96 hours. Our results indicate that approximately 85% of insured windstorm losses in France are attributable to clustered storms.
Building on these findings, we develop vulnerability curves for residential properties that explicitly account for temporally compounded storm events. These curves provide a more realistic representation of windstorm risk than traditional approaches, which typically assess losses either at the scale of individual storms or over an entire winter season. Our results highlight the importance of treating storm clusters as combinations of interdependent events.
How to cite: Hasbini, L., Pascal, Y., Quentin, H., and Simon, B.: Vulnerability curves for clusters of storms - A case study for Generali France, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19428, https://doi.org/10.5194/egusphere-egu26-19428, 2026.
Precipitation is the primary driver of flood risk in France, with both cumulative totals and extreme intensity governing runoff and overflow events. Given the variety of available precipitation products, the choice of data source represents a critical methodological challenge for assessing flood risk. This study evaluates the reliability and predictive sensitivity of several daily precipitation datasets over French territory, including the new SIM2 chain, Météo-France station observations, ECMWF reanalyses (ERA5-Land and ERA-OBS), and regional reanalyses (CERRA and CERRA-Land).
We first perform an in-depth statistical intercomparison for the 1991-2020 period, using the Météo-France station network and ERA-OBS as references. Beyond classic performance metrics (Kling-Gupta Efficiency, RMSE), we place particular emphasis on extreme events using indices such as the Critical Success Index (CSI). Our results identify SIM2 as the most robust overall performer, while ERA-OBS shows high consistency in representing intense rainfall episodes.
Building on this comparison, we assess the operational impact of these data sources through a flood modelling application. Using municipal 'natural disaster' decrees (CatNat) available since 1989, an automatic and fully standardised procedure for variable construction, selection, and modelling is implemented, in which only the precipitation data source varies. We test several machine learning methods (Random Forest, XGBoost etc.) and design variables in multiple formats. This cross-sectional approach reveals how specific biases in meteorological products propagate into flood occurrence predictions. Our findings reinforce the importance of data set selection in hydrometeorological studies and provide a quantitative framework to evaluate the relevance of precipitation sources for the evaluation of insurance-related flood risk in France.
How to cite: Baton, F. and Moriah, M.: From rainfall datasets to flood prediction: evaluating the impact of precipitation data source on catastrophic risk assessment by machine learning in France, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19601, https://doi.org/10.5194/egusphere-egu26-19601, 2026.
With its global footprint, AXA is exposed to multiple natural hazards across the globe. Assessing the frequency and intensity of these events, especially unobserved extremes, is crucial to monitor, mitigate and adapt to the risk they pose.
Tropical cyclones are one of the most scrutinized natural risks by global (re)insurers. Curated observational records date back to the mid-1800s, with increased reliability from the satellite era onwards (post 1970). They are a global risk, with temporal and spatial dependencies between tropical basins. The extreme damage they cause has been at the root of the development of Natural Catastrophe (NATCAT) modelling capabilities by specialized modelling firms, brokers, and (re)insurers.
However, as exposure is increasing and climate is changing, especially in tropical cyclone prone coastal areas globally, the need for robust and accurate estimates of the frequency and intensity of adverse impacts from tropical cyclones is expanding. Observational tropical cyclones datasets like IBTrACS are too short to obtain reliable statistics on rarest and most impactful events.Fine resolution numerical weather models are too computationally expensive to run on extended periods of time.
AI-based weather models running on GPU-accelerated compute infrastructure provide the necessary speedup while maintaining physical accuracy, enabling the generation of thousands of synthetic tropical cyclone seasons. Using NVIDIA's Earth-2 platform, we build a pipeline to produce hundreds of downscaled large ensemble predictions.
This study investigates the potential of these downscaled runs to generate large sets of tropical cyclones physically consistent in space, time and intensity, yielding robust estimates of their impact probability, especially for the rarest events.
How to cite: Rakotoarimanga, H., Meynadier, R., Renard, X., Chalumeau, N., Koch, M., Mustafa, R., and Detyniecki, M.: Estimation of extreme tropical cyclone risk using AI-weather models, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20264, https://doi.org/10.5194/egusphere-egu26-20264, 2026.
Bias-correction methods represent a key processing step in the production of climate indicators derived from climate projections, aiming to reduce systematic model errors and enhance the usability of climate simulations. However, many studies have demonstrated that commonly used bias-correction approaches may introduce important inconsistencies. These include alteration of observed historical estimates, modification or even reversal of the climate change signal projected by climate models, changes in the model uncertainty spread, and strong sensitivity of method performance to the considered variable, climate indicator, region and observational reference dataset. These limitations highlight the risks of applying bias-correction techniques blindly, without careful examination of their implications for each specific case. This contrasts, however, with the strong need for a consistent and comprehensive provision of diverse climate indicators globally to support climate information needs across sectors and stakeholders.
Here, we propose a simple but consistent and accurate delta-based approach for computing adjusted climate indicators, the Indicator Delta Scaling (IDS). The method relies on two basic principles: historical estimates are derived exclusively from observational datasets, while future corrected indicators are obtained by simply updating the observational reference with the projected raw change signal. The method is evaluated globally using CMIP6 historical simulations against observations, which are used both as the historical reference and as a pseudo-future framework. A diverse set of simple, complex, and multivariate climate indicators is used to evaluate the performance of IDS in comparison with state-of-the-art bias-correction approaches, such as Quantile Delta Mapping and the ISIMIP3b method.
Results show that IDS outperforms existing bias-correction methods across multiple evaluation levels. In contrast to other methods, IDS ensures by construction a perfect representation of observed historical estimates, a strict preservation of the modelled delta change and a solid consistency across variables, indicators, and datasets. At the same time, it provides a similar but slightly more accurate estimate of most indicators for future periods. Moreover and importantly, by avoiding the bias correction of input variables' full data distribution, the approach delivers major computational efficiency gains when computing climate indicators.
In summary, the IDS provides a clear, consistent, accurate, and efficient framework for generating ready-to-use climate indicators, addressing key limitations of current bias-correction practices and supporting robust and comprehensive climate risk assessments. The method has been developed within a Copernicus Climate Change Service contract to streamline the global computation of indicators for assessing EU Taxonomy hazards, following the guidance of the European Investment Bank (EIB) for financial risk assessments.
How to cite: Peña-Izquierdo, J., Hofmann, S., Estella, V., Ray, T., Colledge, F., Samantha, L., Steven, W., and Cagnazzo, C.: Indicator Delta Scaling (IDS): A Consistent and Efficient Method for Bias-Correcting Climate Risk Indicators, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-20931, https://doi.org/10.5194/egusphere-egu26-20931, 2026.
Over the past 30+ years, Moody’s/RMS has been at the forefront of catastrophe modelling, developing and supporting models for the global (re)insurance market. Those offerings bring together carefully calibrated stochastic simulations of extreme events with detailed assessments of the vulnerability of a wide range of assets, covering a wide range of perils over key insurance markets. The models, designed to fully capture the risk from today’s climate, have been validated against extensive geophysical observations and against hundreds of billions of dollars of claims data. As part of our offering, and using an extension of the same framework, we also provide for many of those models a view of future risk for a range of scenarios under climate change.
In this presentation, after a general overview of our climate change conditioning framework, we will focus on the specific case of Australian bushfire, a peril which has recently generated a lot of interest in the (re)insurance industry given the large number of recent headline-grabbing events. We will discuss how our CMIP6-based climate change hazard perturbations are derived, as well as the implications of our results for the insurance market. We will also put those results in the context of our other climate change-conditioned catastrophe model offerings available globally.
How to cite: Roy, K., Couldrey, M., and Khare, S.: Assessing the Bottom-Up Financial Impacts from Climate Change Using Catastrophe Modeling: A Case Study of Australian Bushfire Risk, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-21057, https://doi.org/10.5194/egusphere-egu26-21057, 2026.
Climate change intensifies multi‑hazard risks that affect ecosystems, societies, and economies. Addressing these interconnected risks requires integrated systems, harmonized data, and cross‑sectoral collaboration. Research infrastructures (RIs) that observe climate‑ and nature‑related processes generate essential data and services for understanding climate risk determinants: hazard, exposure, and vulnerability, yet their potential remains underutilised by financial, banking, and insurance sectors that increasingly face nature‑dependent risks.
IRISCC (Integrated Research Infrastructure Services for Climate Change Risks; www.iriscc.eu) unites leading European Research Infrastructures (Ris) to provide open, standardized climate‑risk data, tools, and services through transnational and virtual access. With nearly 80 partners across natural and social sciences, IRISCC strengthens the scientific foundations for integrated climate‑risk assessment and supports the translation of RI data and tools into risk‑management landscape
We conducted a stakeholder analysis to map the current and emerging climate‑risk service landscape and to assess how IRISCC services connect with academic, industry and decision making sectors. Survey data from IRISCC partners combined with a preliminary mapping of climate‑risk service providers, show that while strong links exist with EU‑level organizations, direct engagement with financial, banking, and insurance sectors is still very limited. This gap is critical: recent assessments by the European Central Bank indicate that around 72% of European companies depend heavily on at least one ecosystem service, underscoring the financial sector’s exposure to nature degradation (Elderson F.2023, Network for Greening the Financial System NGFS, 2022)
Our findings highlight significant opportunities to embed scientific communities more efficiently, to enhance RI usage, harmonized datasets, and analytical tools into multi‑hazard climate‑risk services. Strengthening these connections can support more robust risk detection, prevention, and early‑warning capabilities, particularly for nature‑dependent industries.
This presentation outlines the key findings from stakeholder analysis, identifies gaps in the current service landscape related to climate risks, and open the potential of IRISCC’s services to contribute to the needs of financial and insurance sectors. By fostering new collaborations and co‑created solutions, IRISCC aims to advance a more holistic, interoperable, and science‑based climate‑risk ecosystem in Europe.
IRISCC is funded by the European Union (project number 101131261). Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union. Neither the European Union nor the granting authority can be held responsible for them.
Elderson F. The economy and banks need nature to survive. European Central Bank. Published June 8, 2023. Accessed January 15, 2026. https://www.ecb.europa.eu/press/blog/date/2023/html/ecb.blog230608~5cffb7c349.en.html
Network for Greening the Financial System (NGFS). Nature‑related risks. Published 2022. Accessed January 15, 2026. https://www.ngfs.net/en/what-we-do/nature-related-risks
How to cite: Kauppi, J., Haapanala, P., Brus, M., Nikolaidis, N., Bäck, J. K., Kivekäs, N., Salgado, M., Kutsch, W., Schaap, D. M. A., Steenberg Larsen, K., Petracca Altieri, R., Murberg, L. E., Lund Myhre, C., Korsgaard, K., Virkki, S., and Rinne, J.: Driving climate risk insights in finance and insurance activities sector with research infrastructures and technologies, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-21704, https://doi.org/10.5194/egusphere-egu26-21704, 2026.
Understanding how extratropical cyclones contribute to extreme sea level (ESL) events is essential for assessing long-term coastal hazards. While individual cyclone impacts are well-documented, the role of cyclone clustering—i.e., multiple storms occurring within short time windows—remains underexplored. Here we present a comprehensive assessment of the relationship between cyclone clustering and ESL variability along the North Sea coast from 1940 to 2024.
We construct a dataset of cyclone life cycles using 3-hourly ERA5 reanalysis and identify clustered events based on consistent spatial and temporal proximity criteria. Concurrently, we analyze tide gauge records from stations surrounding the North Sea coast, applying detrending and band-pass filters to remove long-term and tidal signals to isolate storm-driven sea level variations.
Our results show that cyclone clusters predominantly occur in winter and have increased significantly in frequency over the past 85 years. Comparing sea level responses during clustered and non-clustered periods reveals that clustering events are associated with markedly higher positive sea level anomalies. These differences are especially pronounced in the upper extremes, indicating that clustering enhances the risk of compound ESL events beyond what is observed during non-clustered periods.
This work provides novel evidence that cyclone clustering plays a growing role in shaping extreme sea level behavior in the North Sea region. Our results also underscore the need to incorporate clustering metrics into coastal impact assessments, particularly under changing climate conditions.
How to cite: Li, Z.: Extratropical cyclone clustering amplifies extreme sea-level rise around the North Sea, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-22981, https://doi.org/10.5194/egusphere-egu26-22981, 2026.
Assessing windstorm hazard return periods is crucial for the (re)insurance industry due to the large losses these events can cause. Accurately estimating return periods for specific wind gusts is essential. Traditionally numerical model simulations over multiple years of windstorm events are used for this purpose. However these models may contain biases, such as over-calibration to certain periods (e.g. the 1990s) or major loss events (e.g. Daria and Lothar). Return periods from a numerical model are compared to an existing statistical model and differences explored. From these differences, it is possible to adjust the numerical simulation model output to match the known statistical distribution more closely. The adjustment method adheres to the yearly structure of the numerical simulation model output. It is shown to provide a suitable adjustment for a variety of locations, providing a good use case for the (re)insurance industry. The method is flexible, allowing for more simulated years than the numerical model’s output. This method is applicable to most locations within the European domain, particularly in areas more exposed to extratropical cyclones.
How to cite: Bannister, D., Jones, T., Rye, C., Boyd, J., Stephenson, D., and Priestley, M.: Enhancing European windstorm return period estimates for (re)insurance, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1450, https://doi.org/10.5194/egusphere-egu26-1450, 2026.
European windstorms drive much of the region’s extreme weather, causing catastrophic winds and flooding.
Beyond individual hazards, sequences of windstorms, so-called storm clustering, can make landfall along European coasts and propagate inland, inflicting and compounding socioeconomic impacts. This is directly relevant to local recovery and to understanding how impacts accumulate over short timescales. While several studies have examined how storm intensity may change under future climate conditions, far less attention has been paid to storm clustering, the intensity of clustered storms, and the associated risk.
In this study, we use 2,000 years of climate simulations performed with CESM under present-day and future conditions (100 integrations for 1991–2000 and another 100 for 2091–2100, based on the CMIP5 RCP8.5 scenario) to identify and quantify socioeconomic impacts in Western Europe from extreme winds and their clusters. This large sample provides more robust statistics for detecting sub-monthly clustered storms.
Our objectives are twofold: first, to analyse the physical characteristics of storm clusters; and second, to quantify their socioeconomic relevance in terms of risk and impacts.
How to cite: Meynadier, R., Flaounas, E., Rakotoarimanga, H., Mustafa, R., and Wernli, H.: Current and Future Risks of Storm Clustering in Western Europe., EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-18009, https://doi.org/10.5194/egusphere-egu26-18009, 2026.
