<div>Climate change is increasingly recognized as a top global threat impacting human, environmental, and economic systems.</div><div>When it comes to property & casualty insurance risk, AXA considers every aspect of the risk equation.  As far as natural catastrophes risks are concerned, climate risk is a function of (i) the physical hazard (the severity and frequency of events); (ii) the exposure (the monetary value of insured asset(s)) and (iii) the vulnerability (the susceptibility or damageability of the insured asset(s) to a given hazard intensity). Each of these elements plays a unique role in driving climate risk both now and into the future. The changes AXA see in its year-on-year losses from climate-linked hazards are a function of all risks components and not just the hazard, which is a common misconception.</div><div>AXA is developing internally Natural Hazard models (or Natural Catastrophe models) to estimate the climate risk damages and losses to individual risks or (re)insured portfolios.</div><div>To perform forward looking analysis, AXA identified different complementary approaches that could be envisaged to assess the future of natural hazards risks according to both the peril and region combinations AXA has exposure to as the availability and quality of data for the three drivers of the risk. Those approaches have been notably used for several regulatory climate stress tests AXA was involved in.</div><div>One of them is a global proportional approach simple to implement to consider at a large scale the evolution of hazard, exposure, vulnerability impacts on climate risk for long term time horizons and several warming scenarios. The model is built on current science knowledge related to climate change.</div><div>A more sophisticated approach for local-scale assessment is currently being developed. It consists in integrating in the Natural Hazard models a modified view of hazard (stochastic events catalogue) / exposure / vulnerability capturing forward-looking scenarios. AXA is currently upgrading all its Natural Hazard models in that direction. AXA Europe Flood risk model is for instance assessing future of fluvial and pluvial risks from modified precipitation datasets representative of future warming scenarios. Future precipitation are made of baseline current precipitation from which a “delta” precipitation is added, based on CMIP6 temperature anomalies and Clausius-Clapeyron scaling. Hydrological and hydraulic models are then run using this new future precipitation datasets to generate stochastic flood events catalogue for future warming scenarios risks assessment.</div><div> </div><div> </div>

<p>AXA proposes a novel continental-scale generator of synthetic gridded rainfall daily timeseries (10km resolution) with applications to cross-country risk assessment under current and future climate scenarios. Europe serves as a case-study to demonstrate and assess its performance in terms of hazard modelling and extrapolation to unobserved extreme local and regional events. This generator belongs to the class of time and space reshuffling Stochastic Weather Generators (SWGs) and generates unobserved events by re-sequencing historical multisite timeseries (E-OBS). Consistency at continental scale is ensured by relying on weather regimes and atmospheric situations characterized from the ERA5 reanalysis over Europe. The use of atmospheric drivers and dry-wet alternating cycles allows for the determination of both precipitation-prone situations or on the contrary drier spells, while preserving the physics of the atmospheric water cycle. Spatial reshuffling is introduced by regional differentiation. Transitions between regimes can be either calibrated from the historical data or extrapolated to represent future states of the climate along with an appropriate uplifting of the humidity-related variables. This generator is operationally used at AXA as part of a European flood risk model and serves as the main input to an hydrological and hydraulic model.</p>

No abstract

<p>To understand continental scale flood risks, including spatial and temporal coherence and cascading events, is of particular importance to the insurance industry. For this industry, an “event” entails a certain regulatory duration, and encompasses the spatial scale of the portfolio of the insurer. This requires a large catalogue of statistically well-sampled, climatologically realistic possible events, much longer than any historical record can provide. We hypothesize that events that might have occurred in the recent past, but did not occur, may be generated from shorter duration historical samples, by temporal resampling, and spatial reshuffling.</p><p>In this contribution, we present a model framework – developed by a consortium of Fathom, Deltares, and AXA – that can efficiently compute very large event sets, using synthetically sampled weather (up to many thousands of years) that simulates continuous daily weather and sub-daily (for small-scale pluvial flooding) weather statistics, a gridded hydrological model forced by the synthetic weather that produces long-term hydrological statistics, and a subcatchment-scale fluvial and pluvial flood model archive, produced from large amounts of simulations with the Fathom flood model engine. The framework is setup such that components within the framework can be easily improved or replaced by new components, e.g. providing updated historical baselines for weather generation, enhanced weather generation, enhanced flood maps, or improved hydrological relationships. We present our first simulations using a k-nearest-neighbour weather resampling, using Self-Organizing-Maps, 10,000 years of simulated weather and hydrology, and sampled flood statistics. In forthcoming work, we will improve weather generation mechanism by relaxing the spatial locations of weather systems, and implement climate change.</p>