Windstorm Ciarán: New Enhanced Event Response Benefiting from High-Definition Models
Callum HigginsDecember 04, 2023
Between November 1-2, a rapidly deepening area of low pressure, named Storm Ciarán by the U.K. Met Office and known as Storm Emir in parts of Europe, tracked across northwest Europe and brought strong wind gusts to several countries including France, the Channel Islands, Belgium, the United Kingdom, and the Netherlands.
Ciarán caused widespread damage and disruption across northwest Europe, with northern France, the Channel Islands, Belgium, and the southern U.K. the worst affected.
This has enabled customers with access to ExposureIQ to visualize events against their exposure and run accumulation analyses to identify their exposed limit to the event across their business.
With the release of the Europe Windstorm HD Models on the Risk Modeler™ application in March 2023, including footprint functionality, Moody’s RMS has expanded the breadth of support it can provide for customers when responding to windstorm events like Ciarán.
In addition to the release of the footprint within ExposureIQ and the release of stochastic event selections, Moody’s RMS now also releases footprint reconstructions for European windstorm events that reach our thresholds for loss modeling, available to be run as part of a footprint analysis on Risk Modeler.
These footprints are consistent with the accumulation footprints in ExposureIQ banded by wind speed but are generated for use with the Europe Windstorm HD models, taking advantage of HD modeling methodologies and accounting for the vulnerability of each location to enable event losses to be estimated more accurately.
Recommended Event View
Our recommended view of an event comes from the use of footprint reconstructions to estimate losses, and this is also the methodology we use to generate our industry loss estimates.
There are three reasons why Moody’s RMS recommends using a footprint reconstruction where available.
First, a footprint reconstruction is made bespoke for the event it is generated for, considering as many different aspects of the event as possible, including real-time observations and damage information.
Because of this, the footprints are better able to represent the unique nature of an event, such as the combination of meteorological features of a windstorm, than alternative approaches.
Ciarán is a good example of this. Wind hazard from the event extended far from the storm track – into southern France and even Corsica.
This was due to a complex mixture of meteorological factors including an upper-level trough feature that brought peak wind gusts to southwestern France and a low-pressure feature over northern Italy that produced localized wind gusts across southeastern France and the island of Corsica.
Reconstructing the event ensured that these distinct patterns of wind hazard (as seen in Figure One above) were appropriately represented, flowing through into more accurate estimates of the loss.
Additionally, using footprint reconstructions avoids the problem of a so-called ‘clash’ whereby a stochastic event may provide a good representation of the event in the region of interest, but it also contains wind hazard from unaffected regions.
This can cause issues when these stochastic events are applied to results based on exposure that includes these unaffected regions, with losses out of scope for the event inadvertently incorporated.
Finally, the footprint reconstruction methodology is essentially the same as that used for the historical events within the windstorm models, which were then used to build its calibration target.
Furthermore, Risk Modeler will use the same financial terms of a regular portfolio and the same vulnerability curves of the Europe Windstorm HD Models, including customized vulnerability curves generated via Model Composer.
This ensures full consistency between any loss estimate generated using event response footprints and the model itself.
Realistic Estimates of Claims Volume and Average Claims Size
A more realistic estimate of claims volumes and average claims size can be output using the innovative simulation methodology employed for HD models when compared to non-HD models.
As HD models also account for the probability of zero damage as part of damage sampling, by running event response footprints in Risk Modeler with multiple samples, customers can estimate the probability of a claim as well as the average claim size by location.
When responding to an event, this helps customers triage locations they need to focus on for outreach.
In addition, by upwardly aggregating this information, the likely number of claims by geographical region can be identified, enabling more accurate estimates for how many claims adjustors may be needed to respond as well as the most effective areas to deploy them.
Improved Representation of the Event Through HD Stochastic Event Selections
Leveraging reconstructed wind hazard footprints on Risk Modeler is the recommended approach when modeling losses from European windstorm events, but there are many cases where it may not be possible for our customers to use these.
This could be due to only having access to model results and not the underlying exposure, or when it is too logistically challenging and time-consuming to run footprints against many portfolios – including applying all the necessary financial terms given the time constraints to respond.
However, even where it is not possible to use footprint reconstructions of an event, the HD model offers enhancements to event response from a stochastic event selection perspective, which allows customers to make use of previous analyses.
With a larger number of events represented within the stochastic set and at a higher hazard resolution, it is much easier to identify events that closely reflect the event of interest within the HD model stochastic event set compared to non-HD models.
This is particularly the case for Europe-wide selections designed for customers unable to separate their exposure by country to avoid clash issues as much as possible.
Loss Estimates and Hazard Uncertainty
The capabilities of the HD framework allowed Moody’s RMS Event Response team to take full advantage of the 15 years of experience in responding to events across several peril regions.
Because of this, Moody’s RMS was able to release an Industry Loss Estimate for Windstorm Ciarán/Emir ranging from 0.9 billion euros to 1.5 billion euros for Europe, with 90 percent of the losses in France.
Given the novelty and the flexibility of the HD framework, Moody’s RMS also released two additional footprints, generated by perturbing the wind gusts of the original footprint.
In summary, for the Europe Windstorm HD models, clients can explore secondary uncertainty by changing the number of samples and can evaluate the impact of hazard uncertainty by analyzing the perturbed footprints.
These enhanced event response capabilities highlight the flexibility and the strengths of the HD framework and of the Europe Windstorm HD models, which during the first few months since launch in March 2023 are already widely used in support of the ongoing 1/1 2024 reinsurance renewals.
Ongoing Support for Europe Windstorm Event Response and Further Information
Moody’s RMS continues to monitor European windstorm events this season and stands ready to respond to these as for Windstorm Ciarán.
Callum is the product manager for Moody’s RMS Event Response Services (including HWind) and Agricultural Models and is based in London. Most recently he has been focused on improving client workflows through the integration of event response functionality within the Intelligent Risk Platform.
Previously, Callum has worked on climate change initiatives at Moody’s RMS as well as the 2018 update to the Australia Cyclone Model.
Callum is a Certified Catastrophe Risk Analyst and holds an integrated master’s degree (MEarthSci) in Earth Sciences from Oxford University.