European Windstorms Research Articles

Insurance loss model vs. meteorological loss index – how comparable are their loss estimates for European windstorms?

Abstract. Windstorms affecting Europe are among the natural hazards with the largest socio-economic impacts. Therefore, many sectors like society, the economy, or the insurance industry are highly interested in reliable information on associated impacts and losses. In this study, we compare – for the first time – estimated windstorm losses using a simplified meteorological loss index (LI) with losses obtained from a complex insurance loss (catastrophe) model, namely the European Windstorm Model of Aon Impact Forecasting. To test the sensitivity of LI to different meteorological input data, we furthermore contrast LI based on the reanalysis dataset ERA5 and its predecessor ERA-Interim. We focus on similarities and differences between the datasets in terms of loss values and storm rank for specific historical storm events in the common reanalysis period across 11 European countries. Our results reveal higher LI values for ERA5 than for ERA-Interim for all of Europe (by roughly a factor of 10), coming mostly from the higher spatial resolution in ERA5. The storm ranking is comparable for western and central European countries for both reanalyses, confirmed by high correlation values between 0.6 and 0.89. Compared to the Aon Impact Forecasting model, LI ERA5 shows comparable storm ranks, with correlation values ranging between 0.45 and 0.8. In terms of normalized loss, LI exhibits overall lower values and smaller regional differences. Compared to the market perspective represented by the insurance loss model, LI seems to have particular difficulty in distinguishing between high-impact events at the tail of the wind gust distribution and moderate-impact events. Thus, the loss distribution in LI is likely not steep enough, and the tail is probably underestimated. Nevertheless, it is an effective index that is suitable for estimating the impacts of storm events and ranking storm events, precisely because of its simplicity.

Do AI models produce better weather forecasts than physics-based models? A quantitative evaluation case study of Storm Ciarán

There has been huge recent interest in the potential of making operational weather forecasts using machine learning techniques. As they become a part of the weather forecasting toolbox, there is a pressing need to understand how well current machine learning models can simulate high-impact weather events. We compare short to medium-range forecasts of Storm Ciarán, a European windstorm that caused sixteen deaths and extensive damage in Northern Europe, made by machine learning and numerical weather prediction models. The four machine learning models considered (FourCastNet, Pangu-Weather, GraphCast and FourCastNet-v2) produce forecasts that accurately capture the synoptic-scale structure of the cyclone including the position of the cloud head, shape of the warm sector and location of the warm conveyor belt jet, and the large-scale dynamical drivers important for the rapid storm development such as the position of the storm relative to the upper-level jet exit. However, their ability to resolve the more detailed structures important for issuing weather warnings is more mixed. All of the machine learning models underestimate the peak amplitude of winds associated with the storm, only some machine learning models resolve the warm core seclusion and none of the machine learning models capture the sharp bent-back warm frontal gradient. Our study shows there is a great deal about the performance and properties of machine learning weather forecasts that can be derived from case studies of high-impact weather events such as Storm Ciarán.

Return levels of extreme European windstorms, their dependency on the North Atlantic Oscillation, and potential future risks

Abstract. Windstorms are the most damaging natural hazard across western Europe. Risk modellers are limited by the observational data record to only ∼ 60 years of comprehensive reanalysis data that are dominated by considerable inter-annual variability. This makes estimating return periods of rare events difficult and sensitive to the choice of the historical period used. This study proposes a novel statistical method for estimating wind gusts across Europe based on observed windstorm footprints. A good description of extreme wind speeds is obtained by assuming that gust speed peaks over threshold are distributed exponentially, i.e. a generalised Pareto distribution having a zero shape parameter. The threshold and tail scale parameter are estimated at each location and used to calculate estimates of the 10- and 200-year return levels. The North Atlantic Oscillation (NAO) is particularly important for modulating lower return levels and modulating the threshold, with a less detectable influence on rarer extremes and the tail scale parameter. The length of historical data required to have the lowest error in estimating return levels is quantified using both observed and simulated time series of the historical NAO. For reducing errors in estimating 200-year return levels of an independent 10-year period, a data catalogue of at least 20 years is required. For lower return levels the NAO has a stronger influence on estimated return levels, and so there is more variability in estimates. Using theoretical estimates of future NAO states, return levels are largely outside the historical uncertainty, indicating significant increases in risk potential from windstorms in the next 100 years. Our method presents a framework for assessing high-return-period events across a range of hazards without the additional complexities of a full catastrophe model.

A long record of European windstorm losses and its comparison to standard climate indices

Abstract. Traditional insurance has both a great exposure to decadal variations in European storm activity and the ability to adjust its business strategy over these timescales. Hence, the recent development of skilful predictions of multiannual mean European winter climate seems a very welcome addition to the long list of ways that researchers have improved management of windstorm risk. Yet companies do not use these forecasts of mean winter climate to adjust their view of risk. The main reason is the lack of a long, reliable record of losses to understand how forecasted time-mean circulation anomalies relate to the damage from a few, intense storms. This study fills that gap with a European windstorm loss record from 1950 to 2022, based on ERA5 peak near-surface winds per event which were converted to losses using an established damage function. The resulting dataset successfully identifies major storms over the past 70 years and simulates the multidecadal variations from low values in the 1960s up to high levels in the 1980s and 1990s then down to the 2010s. However, it underestimated the steepness of the observed loss decline from the stormy end of the 20th century to the lull over the past 20 years. This was caused by a quite flat trend in ERA5 extreme winds over the period, in contrast to the significant decline in observed peak gusts. Imposing these gust trends on ERA5 peak winds reconciled modelled losses with industry experience over the past few decades. Indices of European winter climate used in long-range forecasting were compared to the new modelled loss dataset. They had correlations of around 0.4 at interannual timescales, rising to about 0.7 for decadal and longer variations. Notably, the climate indices have a similar multidecadal trend as ERA5 extreme winds in modern times, including a less steep decline than found in observed gusts and losses. Further investigation of the modern-day divergence between climate indices and losses may help connect decadal forecasting to insurance.

Future increased risk from extratropical windstorms in northern Europe

European windstorms cause socioeconomic losses due to wind damage. Projections of future losses from such storms are subject to uncertainties from the frequency and tracks of the storms, their intensities and definitions thereof, and socio-economic scenarios. We use two storm severity indices applied to objectively identified extratropical cyclone footprints from a multi-model ensemble of state-of-the-art climate models under different future socio-economic scenarios. Here we show storm frequency increases across northern and central Europe, where the meteorological storm severity index more than doubles. The population-weighted storm severity index more than triples, due to projected population increases. Adapting to the increasing wind speeds using future damage thresholds, the population weighted storm severity index increases are only partially offset, despite a reduction in the meteorological storm severity through adaptation. Through following lower emissions scenarios, the future increase in risk is reduced, with the population-weighted storm severity index increase more than halved.

Multiple Large‐Scale Dynamical Pathways for Pan–Atlantic Compound Cold and Windy Extremes

AbstractWinter cold spells over North America have been correlated with European wind extremes, but the physical mechanisms behind such “pan‐Atlantic” compound extremes have not been clarified yet. In this study, we propose that pan–Atlantic cold and windy extremes occur following two possible dynamical pathways. The first one involves the propagation of a Rossby wave train from the Pacific Ocean, associated with windstorms over north‐western Europe in the 5–10 days after the cold spell peak. The second is associated with a high‐latitude anticyclone over the North Atlantic and an equatorward‐shifted jet, leading to windstorms over south‐western Europe already in the days preceding the cold spell peak. European windstorms are thus systematically tied to North American cold spells according to the different flow configuration. The analysis underscores that apparently similar surface extremes might be driven by different processes, and conflating them together could lead to misleading conclusions.

Using high-resolution global climate models from the PRIMAVERA project to create a European winter windstorm event set

Abstract. PRIMAVERA (process-based climate simulation: advances in high-resolution modelling and European climate risk assessments) was a European Union Horizon 2020 project whose primary aim was to generate advanced and well-evaluated high-resolution global climate model datasets for the benefit of governments, business and society in general. Following consultation with members of the insurance industry, we have used a PRIMAVERA multi-model ensemble to generate a European winter windstorm event set for use in insurance risk analysis, containing approximately 1300 years of windstorm data. The data are available at https://doi.org/10.5281/zenodo.6492182. To create the storm footprints for the event set, the storms in the PRIMAVERA models are identified through tracking. A method is developed to separate the winds from storms occurring in the domain at the same time. The wind footprints are bias corrected and converted to 3 s gusts onto a uniform grid using quantile mapping. The distribution of the number of model storms per season as a function of estimated loss is consistent with re-analysis, as are the total losses per season, and the additional event set data greatly reduce uncertainty on return period magnitudes. The event set also reproduces the temporally clustered nature of European windstorms. Since the event set is generated from global climate models, it can help to quantify the non-linear relationship between large-scale climate indices such as the North Atlantic Oscillation (NAO) and windstorm damage. Although we find only a moderate positive correlation between extended winter NAO and storm damage in northern European countries (consistent with re-analysis), there is a large change in risk of extreme seasons between negative and positive NAO states. The intensities of the most severe storms in the event set are, however, sensitive to the gust conversion and bias correction method used, so care should be taken when interpreting the expected damages for very long return periods.

Functional Peaks-Over-Threshold Analysis

AbstractPeaks-over-threshold analysis using the generalised Pareto distribution is widely applied in modelling tails of univariate random variables, but much information may be lost when complex extreme events are studied using univariate results. In this paper, we extend peaks-over-threshold analysis to extremes of functional data. Threshold exceedances defined using a functional r are modelled by the generalised r-Pareto process, a functional generalisation of the generalised Pareto distribution that covers the three classical regimes for the decay of tail probabilities, and that is the only possible continuous limit for r-exceedances of a properly rescaled process. We give construction rules, simulation algorithms and inference procedures for generalised r-Pareto processes, discuss model validation and apply the new methodology to extreme European windstorms and heavy spatial rainfall.

European Windstorm Risk of Post‐Tropical Cyclones and the Impact of Climate Change

AbstractSainsbury et al. (2020, https://doi.org/10.1029/2020GL089853) highlighted the fact that although post‐tropical cyclones only consist of a small part of the cyclones impacting Northern Europe (∼1%) during the hurricane season, their contribution to storms with at least storm force (>25 m/s) is about 10 times larger and therefore pose a substantial windstorm risk for Northern Europe. Overall post‐TCs impacting Northern Europe are significantly more intense than midlatitude cyclones. This is related to their combined tropical (warm seawater) and midlatitude (baroclinic instability) energy source. A few studies have suggested that in a warmer climate the amount and intensity of post‐tropical cyclones might increase due to stronger tropical cyclones (TCs), warmer oceans, and pole and eastward extension of TC genesis regions. This would further enhance the windstorm risk of post‐TCs. Further high‐resolution modeling research is required to delineate this risk.

On the Dependency of Atlantic Hurricane and European Windstorm Hazards

AbstractThe Atlantic hurricane season and the European windstorm season are found to be negatively correlated in a seasonal forecast model. The probability of extremes occurring in both seasons is compared to the probability of extremes in each season being independent of one another. An above average Atlantic hurricane season is followed by an above average European windstorm season less often than if they were independent, consistent across three intensity measures. The El Niño–Southern Oscillation is found to be in the positive (negative) phase when hurricane activity is suppressed (enhanced) and European windstorm activity is enhanced (suppressed). A clear extratropical response in the seasonal forecast model to El Niño/La Niña provides a probable pathway for the observed correlation between the extreme event seasons. This result has important predictability implications for both the actuarial and seasonal forecasting communities.

Modelling The Spatial Extent and Severity of Extreme European Windstorms

Summary Windstorms are a primary natural hazard affecting Europe that are commonly linked to substantial property and infrastructural damage and are responsible for the largest spatially aggregated financial losses. Such extreme winds are typically generated by extratropical cyclone systems originating in the North Atlantic and passing over Europe. Previous statistical studies tend to model extreme winds at a given set of sites, corresponding to inference in an Eulerian framework. Such inference cannot incorporate knowledge of the life cycle and progression of extratropical cyclones across the region and is forced to make restrictive assumptions about the extremal dependence structure. We take an entirely different approach which overcomes these limitations by working in a Lagrangian framework. Specifically, we model the development of windstorms over time, preserving the physical characteristics linking the windstorm and the cyclone track, the path of local vorticity maxima, and make a key finding that the spatial extent of extratropical windstorms becomes more localized as its magnitude increases irrespective of the location of the storm track. Our model allows simulation of synthetic windstorm events to derive the joint distributional features over any set of sites giving physically consistent extrapolations to rarer events. From such simulations improved estimates of this hazard can be achieved in terms of both intensity and area affected.

Human influence on European winter wind storms such as those of January 2018

Abstract. Several major storms pounded western Europe in January 2018, generating large damages and casualties. The two most impactful ones, Eleanor and Friederike, are analysed here in the context of climate change. Near surface wind speed station observations exhibit a decreasing trend in the frequency of strong winds associated with such storms. High-resolution regional climate models, on the other hand, show no trend up to now and a small increase in storminess in future due to climate change. This shows that factors other than climate change, which are not in the climate models, caused the observed decline in storminess over land. A large part is probably due to increases in surface roughness, as shown for a small set of stations covering the Netherlands and in previous studies. This observed trend could therefore be independent from climate evolution. We concluded that human-induced climate change has had so far no significant influence on storms like the two mentioned. However, all simulations indicate that global warming could lead to a marginal increase (0 %–20 %) in the probability of extreme hourly winds until the middle of the century, consistent with previous modelling studies. This excludes other factors, such as surface roughness, aerosols, and decadal variability, which have up to now caused a much larger negative trend. Until these factors are correctly simulated by climate models, we cannot give credible projections of future storminess over land in Europe.

Selective ensemble mean technique for severe European windstorms

We show that ensemble forecasts of extreme European windstorms can be improved up to a lead time of 36–48 hr by sub‐selecting ensemble members based on their performance at very short lead times (up to 12 hr). This applies to both the ensemble mean position of the cyclone centre and the ensemble windstorm footprint over the continent. A number of ensemble forecasts, including those from the ECMWF Ensemble Prediction System, are initialised every 12 hr and disseminated several hours after initialisation; therefore our approach has the potential to provide improved forecasts in an operational context. The analysis is performed on GEFS reforecast data.

The role of serial European windstorm clustering for extreme seasonal losses as determined from multi-centennial simulations of high-resolution global climate model data

Abstract. Extratropical cyclones are the most damaging natural hazard to affect western Europe. Serial clustering occurs when many intense cyclones affect one specific geographic region in a short period of time which can potentially lead to very large seasonal losses. Previous studies have shown that intense cyclones may be more likely to cluster than less intense cyclones. We revisit this topic using a high-resolution climate model with the aim to determine how important clustering is for windstorm-related losses. The role of windstorm clustering is investigated using a quantifiable metric (storm severity index, SSI) that is based on near-surface meteorological variables (10 m wind speed) and is a good proxy for losses. The SSI is used to convert a wind footprint into losses for individual windstorms or seasons. 918 years of a present-day ensemble of coupled climate model simulations from the High-Resolution Global Environment Model (HiGEM) are compared to ERA-Interim reanalysis. HiGEM is able to successfully reproduce the wintertime North Atlantic/European circulation, and represent the large-scale circulation associated with the serial clustering of European windstorms. We use two measures to identify any changes in the contribution of clustering to the seasonal windstorm loss as a function of return period. Above a return period of 3 years, the accumulated seasonal loss from HiGEM is up to 20 % larger than the accumulated seasonal loss from a set of random resamples of the HiGEM data. Seasonal losses are increased by 10 %–20 % relative to randomized seasonal losses at a return period of 200 years. The contribution of the single largest event in a season to the accumulated seasonal loss does not change with return period, generally ranging between 25 % and 50 %. Given the realistic dynamical representation of cyclone clustering in HiGEM, and comparable statistics to ERA-Interim, we conclude that our estimation of clustering and its dependence on the return period will be useful for informing the development of risk models for European windstorms, particularly for longer return periods.

Quantification of extremal dependence in spatial natural hazard footprints: independence of windstorm gust speeds and its impact on aggregate losses

Abstract. Natural hazards, such as European windstorms, have widespread effects that result in insured losses at multiple locations throughout a continent. Multivariate extreme-value statistical models for such environmental phenomena must therefore accommodate very high dimensional spatial data, as well as correctly representing dependence in the extremes to ensure accurate estimation of these losses. Ideally one would employ a flexible model, able to characterise all forms of extremal dependence. However, such models are restricted to a few dozen dimensions, hence an a priori diagnostic approach must be used to identify the dominant form of extremal dependence. Here, we present various approaches for exploring the dominant extremal dependence class in very high dimensional spatial hazard fields: tail dependency measures, copula fits, and conceptual loss distributions. These approaches are illustrated by application to a data set of high-dimensional historical European windstorm footprints (6103 spatial maps of 3-day maximum gust speeds at 14 872 locations). We find there is little evidence of asymptotic extremal dependency in windstorm footprints. Furthermore, empirical extremal properties and conceptual losses are shown to be well reproduced using Gaussian copulas but not by extremally dependent models such as Gumbel copulas. It is conjectured that the lack of asymptotic dependence is a generic property of turbulent flows. These results open up the possibility of using geostatistical Gaussian process models for fast simulation of windstorm hazard fields.

Increased wind risk from sting-jet windstorms with climate change

Extra-tropical cyclones dominate autumn and winter weather over western Europe. The strongest cyclones, often termed windstorms, have a large socio-economic impact on landfall due to strong surface winds and coastal storm surges. Climate model integrations have predicted a future increase in the frequency of, and potential damage from, European windstorms and yet these integrations cannot properly represent localised jets, such as sting jets, that may significantly enhance damage. Here we present the first prediction of how the climatology of sting-jet-containing cyclones will change in a future warmer climate, considering the North Atlantic and Europe. A proven sting-jet precursor diagnostic is applied to 13 year present-day and future (~2100) climate integrations from the Met Office Unified Model in its Global Atmosphere 3.0 configuration. The present-day climate results are consistent with previously-published results from a reanalysis dataset (with around 32% of cyclones exhibiting the sing-jet precursor), lending credibility to the analysis of the future-climate integration. The proportion of cyclones exhibiting the sting-jet precursor in the future-climate integration increases to 45%. Furthermore, while the proportion of explosively-deepening storms increases only slightly in the future climate, the proportion of those storms with the sting-jet precursor increases by 60%. The European resolved-wind risk associated with explosively-deepening storms containing a sting-jet precursor increases substantially in the future climate; in reality this wind risk is likely to be further enhanced by the release of localised moist instability, unresolved by typical climate models.

Quantifying the extremity of windstorms for regions featuring infrequent events

AbstractThis paper introduces the Distribution‐Independent Storm Severity Index (DI‐SSI). The DI‐SSI represents an approach to quantify the severity of exceptional surface wind speeds of large scale windstorms that is complementary to the SSI introduced by Leckebusch et al. While the SSI approaches the extremeness of a storm from a meteorological and potential loss (impact) perspective, the DI‐SSI defines the severity in a more climatological perspective. The idea is to assign equal index values to wind speeds of the same singularity (e.g. the 99th percentile) under consideration of the shape of the tail of the local wind speed climatology. Especially in regions at the edge of the classical storm track, the DI‐SSI shows more equitable severity estimates, e.g. for the extra‐tropical cyclone Klaus. In order to compare the indices, their relation with the North Atlantic Oscillation is studied, which is one of the main large scale drivers for the intensity of European windstorms.

Generalised additive point process models for natural hazard occurrence

Point processes are a natural class of models for representing occurrences of various types of natural hazard event. Flexibly implementing such models is often hindered by intractable likelihood forms. Consequently, the rates of point processes tend to be reduced to parametric forms, or the processes are discretised to give data of readily modelled “count‐per‐unit” type. This work proposes generalised additive model forms for point process rates. The resulting low‐rank spatiotemporal representations of rates, coupled with the Laplace approximation, makes the restricted likelihood relatively tractable and hence inference for such models possible. The models can also be interpreted from a regression perspective. The proposed models are used to estimate different types of Cox process and then spatiotemporal variation in European windstorms. Through a combination of thin‐plate and cubic regression splines and their tensor product, established relationships between where windstorms occur and the state of the North Atlantic Oscillation are confirmed and then expanded to bring detailed understanding of within‐year variation, which has otherwise not been possible with count‐based models.

Drivers for the deepening of severe European windstorms and their impacts on forecast quality

European windstorms are a high‐impact weather phenomenon, regularly inflicting substantial damages, both human and economic. This study examines a set of objectively selected intense European windstorms from the 1979–2015 period using re‐analysis and forecast products from the European Centre for Medium‐Range Weather Forecasts (ECMWF). The storms are first categorized with respect to their diabatic relative to baroclinic contribution to deepening using the pressure tendency equation, and secondly with respect to their track relative to the jet stream as the large‐scale element controlling storm deepening and propagation. As expected, baroclinic processes dominate the majority of storms, such that deepening is closely related to warm advection ahead of the cyclone centre. Contributions from diabatic processes vary strongly and exceed those from horizontal temperature advection in 10 out of 58 cases, with values of up to 60%. Remarkably, in several cases, planetary waves in the stratosphere appear to facilitate cyclogenesis but can also act to oppose deepening in a few cases. The diabatic contribution is significantly correlated with the time a given storm spends on the equatorward side of the jet, where there is greater potential for diabatic processes in the warm, moist air. In terms of forecast quality and consistent with previous studies, the storm core pressure is generally underestimated and storms tend to be too slow and shifted south in the forecast, particularly for longer lead times. These biases, however, reduce markedly with the improvement of the operational system over time. There is no systematic dependence of forecast behaviour on the diabatic contribution or track relative to the jet. In the future, some of these analyses should be repeated with homogeneous reforecast data to substantiate these findings better.

The 21st century decline in damaging European windstorms

Abstract. A decline in damaging European windstorms has led to a reduction in insured losses in the 21st century. This decline is explored by identifying a damaging windstorm characteristic and investigating how and why this characteristic has changed in recent years. This novel exploration is based on 6103 high-resolution model-generated historical footprints (1979–2014), representing the whole European domain. The footprint of a windstorm is defined as the maximum wind gust speed to occur at a set of spatial locations over the duration of the storm. The area of the footprint exceeding 20 ms−1 over land, A20, is shown to be a good predictor of windstorm damage. This damaging characteristic has decreased in the 21st century, due to a statistically significant decrease in the relative frequency of windstorms exceeding 20 ms−1 in north-western Europe, although an increase is observed in southern Europe. This is explained by a decrease in the quantiles of the footprint wind gust speed distribution above approximately 18 ms−1 at locations in this region. In addition, an increased variability in the number of windstorm events is observed in the 21st century. Much of the change in A20 is explained by the North Atlantic Oscillation (NAO). The correlation between winter total A20 and winter-averaged mean sea-level pressure resembles the NAO pattern, shifted eastwards over Europe, and a strong positive relationship (correlation of 0.715) exists between winter total A20 and winter-averaged NAO. The shifted correlation pattern, however, suggests that other modes of variability may also play a role in the variation in windstorm losses.