Flood risk experts have all got the invitation for the “climate challenge ball”, but do we know what to wear? Is there even a dress code? The insurance sector is no stranger when it comes to dealing with risks with huge uncertainties. We never look at absolute numbers, our party glasses show us the world in distributions of possible events. Yet, we still don’t feel comfortable enough with the uncertainty ranges in climate models, especially when compounded by catastrophe model uncertainties. But we should be, party dresses are never comfortable when you first try them on.
It is time to invest; invest in knowledge to support the greening of the financial system.
We have been equipped with sophisticated catastrophe flood models for some time now. They are getting more and more comprehensive and detailed by the day; capturing various sources of flooding, using high-resolution data, incorporating multi-peril interactions, providing us the full range of possibilities with uncertainty ranges. Now we are asking them to incorporate the insights from climate model outputs on potential futures. Climate change threatens the stability of the financial system, and while catastrophe models are focussed on just the next year or so, we must now look further into the future and assess climate change related risks 10-20 years or more ahead. It is time to invest; invest in knowledge to support the greening of the financial system. Our long term Willis Research Network (WRN) collaborator, Professor of Hydrology and Climate Change Chris Kilsby of Newcastle University, says that most insurance people have come to the climate party quite late, but he was able to catch us up and give us the highlights of the night so far.
During this year’s Flood Club meeting, the fifth of its kind which brings together the flood modelling community to discuss current technical and research questions around flood risk, Professor Kilsby underlined the fact that we need to embrace the challenge of capturing urban flood risk better first. He amiably labels the ‘pluvial flood modelling approach in current catastrophe modelling tools’ as the ‘unhappy Cinderella’; being late and poorly prepared to manage the “climate challenge ball”. Modelling pluvial flood in cities is complicated yet, applying a broad brush approach and simplifying the complexities of urban settings will increase the uncertainties in model outputs. To start with, we need to better understand the present-day risk. Maybe the next step is a hybrid solution of using stochastic catastrophe models with very high-resolution urban flood models, such as CityCAT, to simulate what happens around the buildings and infrastructures well, which is key for the insurance market. With their new UK Flood Model1, Fathom is taking that first step of better capturing the present-day flood risk. Their method uses LIDAR in the terrain model, implementing a better bathymetry calculation, dynamically modelling defences and combining coastal hazard in the ‘inland flood risk view’ which are all part of improving our understanding of present-day flood risk.
With more insight coming from new generation convective permitting models, we are getting a clearer vision on increasing intensities.
When we have a good handle of present-day risk then the next question is, he asks “how do we include the climate model outputs into flood modelling? Do we need to create new models every time UK Met Office creates a new Climate Model output?” He suggests a practical way2; creating a library of storms with various intensities and analyse their shift under the influence of climate change. With more insight coming from new generation convective permitting models, we are getting a clearer vision on increasing intensities. During the conversation, Oliver Wing of Fathom also provided a practical way to incorporate climate change signals through modelling future fluvial and pluvial risk using the ‘change factor approach’ without ignoring the uncertainty ranges coming from climate models.
All the Flood club presenters and the meeting participants highlighted the fact that to get the complete picture of future flood risks, we do need to quantify the changes on hazard, exposure and vulnerability. Professor Kilsby suggested using a consistent ‘urban development model’ for the same time period that the hazard change is analysed, otherwise we have no choice but to make sweeping assumptions about land use change, flood management and urban development, or simply, and unrealistically, assume no changes to exposures and vulnerabilities to flood in the future.
Representing the industry, Cameron Rye of MS Amlin, brought ‘Kermit’ to the “climate challenge ball”. He uses the analogy of the impact of climate change as the ‘boiling frog’. As is a common question in risk assessment, he urges companies to ask ‘what is the worst that could happen and how likely it is’, but from a longer term climate change perspective. Cameron suggests that in addition to the exploratory scenarios, we need to adopt a ‘normative approach’3 when it comes to understanding and managing climate change impact on flood risk, or in any type of catastrophe risk. This is a useful concept and helps different stakeholders in the risk management business communicate on level terms.
The path to fruitful collaboration goes through speaking the same language, or in the case of financial risk assessment, using the same risk metrics. The concept of insurance is built around managing risk, understanding uncertainties and communicating them in terms of financial impacts. Climate change however, puts a spanner in the works of our usual risk landscape, and so we need to bring in the knowledge of climate science from its source; academia, and find a way to appropriately apply this knowledge to our trusted tools; catastrophe models, and finally make sure the outputs are reflected in the same metrics we use every day for decision making across the finance industry. There will be no one-size-fits-all approach, but if the raw materials of the risk assessment are the same, there should at least be some consistency and credibility to results. Keep in mind that each company is unique with their risk appetite and their preferred risk metrics. One of the key questions each company should ask is when and where does flood risk become uninsurable for them under future climate change scenarios. The answer will have an impact on not just the company/industry but also on the government investment plans. One way to go about it is to analyse the change in hazard, exposure and vulnerability components in a holistic catastrophe modelling framework but used to explore different future warmer worlds using outputs from climate model experiments. Climate change impacts are not just classified as physical risks but also fall in to transition risk and liability risk buckets. The specific impact on each company will be different, hence the most useful solutions when it comes to managing climate change impact on your risk profile will be nuanced. At the “climate challenge ball” you don’t want to wear the same dress as everyone else. If you want something that fits your style, and to stand out from the crowd, custom-made is the way to go!
Industry and the research communities are increasingly overlapping through initiatives like the Willis Research Network. The climate challenge will not be solved by any one institution or service, so collaboration is the key that will unlock the best and most useful solutions. Flood Club creates a great platform to facilitate brainstorming opportunities to the re/insurance industry with academic community and model vendors. The languages are becoming more similar and techniques are starting to cross-pollinate to be used in both camps. There is plenty of research on flood risk that needs to be done whether it’s on urban flood modelling or global scale flood modelling, or whether we purely focus on hazards or try to incorporate policy and flood defences investment strategies into future climate risk assessment. Using integrated assessment models (IAMs), to understand different socio-economic pathways and how they impact the insurance and banking industry or finance sector in general are some of the topics both the industry and academia should be focusing on. The WRN will leverage latest research and collaborative opportunities, and make sure that we stay focussed on the credible applications and tangible research outputs that we can bring from the academic community, to address the questions that the industry needs to respond to, in the face of changing climate risks.
1 https://www.fathom.global/fathom-uk
2 https://www.tyndall.ac.uk/OpenCLIM
3 Rye et al. (2021). A. Normative approach to risk management for insurers. Nature Climate Change. https://doi.org/10.1038/s41558-021-01031-8
