Olga Kaiser, Dimitri Igdalov, Olivia Martius, Illia Horenko
Focusing on regression based analysis of extremes in a presence of systematically missing covariates, this work presents a data-driven spatio-temporal regression based clustering of threshold excesses. It is shown that in a presence of systematically missing covariates the behavior of threshold excesses becomes nonstationary and nonhomogenous. The presented approach describes this complex behavior by a set of local stationary Generalized Pareto Distribution (GPD) models, where the parameters are expressed as regression models, and a latent spatio-temporal switching process. The spatio-temporal switching process is resolved by the nonparametric Finite Element Methodology for time series analysis with Bounded Variation of the model parameters (FEM-BV). The presented FEM-BV-GPD approach goes beyond strong a priori assumptions made in standard latent class models like Mixture Models and Hidden Markov Models. In addition, it provides a pragmatic description of the underlying dependency structure. The performance of the framework is demonstrated on historical precipitation data for Switzerland and compared with the results obtained by the standard methods on the same data.
Alice Portal, Andrea Angelidou, Raphael Rousseau-Rizzi, Shira Raveh-Rubin, Yonatan Givon, Jennifer L Catto, Francesco Battaglioli, Mateusz Taszarek, Emmanouil Flaounas, Olivia Martius
Understanding convective processes leading to severe weather hazards within Mediterranean cyclones is relevant for operational forecasters, insurance industry, and enhancing societal preparedness. In this work we examine the climatological link between Mediterranean cyclones and atmospheric conditions conducive to the formation of severe convection and convective hazards (convective precipitation, lightning and hail potential). Using ATDnet lightning detections we find that, from autumn to spring, 20 to 60% of lightning hours over the Mediterranean basin and adjacent land regions are associated with the presence of a nearby cyclone. Based on reanalysis data, severe convective environments, deep, moist convection (i.e., lightning potential) and related hazards are frequent in the warm sector of Mediterranean cyclones and to the north-east of their centres. In agreement with previous literature, convective processes and hazards peak approximately six hours prior to the time of minimum pressure of the cyclone centre. Moreover, severe convective environments are often detected in cyclone categories typical of transition seasons (especially autumn) and summer, while they are rarer in deep baroclinic cyclones with peak occurrence during winter. Finally, we show that dynamical cyclone features distinguish regions favourable to deep, moist convection. Warm conveyor belts of Mediterranean cyclones, characterised by large-scale ascent and located in regions of high thermodynamic instability, have the largest lightning potential. The potential is only half as intense along the cyclones' cold fronts.
Monika Feldmann, Tom Beucler, Milton Gomez, Olivia Martius
Severe convective storms are among the most dangerous weather phenomena and accurate forecasts mitigate their impacts. The recently released suite of AI-based weather models produces medium-range forecasts within seconds, with a skill similar to state-of-the-art operational forecasts for variables on single levels. However, predicting severe thunderstorm environments requires accurate combinations of dynamic and thermodynamic variables and the vertical structure of the atmosphere. Advancing the assessment of AI-models towards process-based evaluations lays the foundation for hazard-driven applications. We assess the forecast skill of three top-performing AI-models for convective parameters at lead-times of up to 10 days against reanalysis and ECMWF's operational numerical weather prediction model IFS. In a case study and seasonal analyses, we see the best performance by GraphCast and Pangu-Weather: these models match or even exceed the performance of IFS for instability and shear. This opens opportunities for fast and inexpensive predictions of severe weather environments.
Martin Aregger, Olivia Martius, Urs Germann, Alessandro Hering
Differential reflectivity columns (ZDRC) have been shown to provide information about a storm's updraft intensity and size. The updraft's characteristics, in turn, influence a severe storm's propensity to produce hail and the size of said hail. Consequently, there is the potential to use ZDRC for the detection and sizing of hail. In this observational study, we investigate the characteristics of ZDRC (volume, height, area, maximum ZDR within) automatically detected on an operational C-band radar network in Switzerland and relate them to hail on the ground using 173'000 crowdsourced hail reports collected over a period of 3.5 years. We implement an adapted version of an established ZDRC detection algorithm on a 3D composite of ZDR data derived from five Swiss weather radars. The composite, in combination with the dense network of radars located on differing altitudes up to 3000 m.a.s.l, helps to counteract the effects of the complex topography of the study region. The alpine region presents visibility and data quality challenges, which are especially crucial for measuring ZDRC. Our analysis finds ZDRC present in most hail-producing storms, with higher frequencies in storms producing severe hail. Further, when looking at lifetime maximum values, we find significant differences in various ZDRC characteristics between hail-producing and non-hail-producing storms. We also attempt to determine thresholds to differentiate between storm types. The temporal evolution of the ZDRC proves challenging to investigate due to their intermittent nature. Nevertheless, the peak values of the ZDRC characteristics are most often measured 5-10 minutes before the first hail reports on the ground, highlighting the potential for ZDRC to be used in warning applications.
Alessio Ciullo, Eric Strobl, Simona Meiler, Olivia Martius, David N. Bresch
Jun 28, 2022·q-fin.RM·PDF Extreme weather events can have severe impacts on national economies, leading the recovery of low- to middle-income countries to become reliant on foreign financial aid. Foreign aid, however, is slow and uncertain. Therefore, the Sendai Framework and the Paris Agreement advocate for more resilient financial instruments like sovereign catastrophe risk pools. Existing pools, however, might not fully exploit financial resilience potentials because they were not designed with the goal of maximizing risk diversification and they pool risk only regionally. To address this, we introduce a method that forms pools maximizing risk diversification and which selects countries with low bilateral correlations or low shares in the pool risk. We apply the method to explore the benefits of global pooling with respect to regional pooling. We find that global pooling increases risk diversification, it lowers countries shares in the pool risk and it increases the number of countries profiting from risk pooling.
Pauline Rivoire, Philomène Le Gall, Anne-Catherine Favre, Philippe Naveau, Olivia Martius
Accurate estimation of daily rainfall return levels associated with large return periods is needed for a number of hydrological planning purposes, including protective infrastructure, dams, and retention basins. This is especially relevant at small spatial scales. The ERA-5 reanalysis product provides seasonal daily precipitation over Europe on a 0.25 x 0.25 grid (about 27 x 27 km). This translates more than 20,000 land grid points and leads to models with a large number of parameters when estimating return levels. To bypass this abundance of parameters, we build on the regional frequency analysis (RFA), a well-known strategy in statistical hydrology. This approach consists in identifying homogeneous regions, by gathering locations with similar distributions of extremes up to a normalizing factor and developing sparse regional models. In particular, we propose a step-by-step blueprint that leverages a recently developed and fast clustering algorithm to infer return level estimates over large spatial domains. This enables us to produce maps of return level estimates of ERA-5 reanalysis daily precipitation over continental Europe for various return periods and seasons. We discuss limitations and practical challenges and also provide a git hub repository. We show that a relatively parsimonious model with only a spatially varying scale parameter can compete well against statistical models of higher complexity.
Sam Allen, Jonas Bhend, Olivia Martius, Johanna Ziegel
To mitigate the impacts associated with adverse weather conditions, meteorological services issue weather warnings to the general public. These warnings rely heavily on forecasts issued by underlying prediction systems. When deciding which prediction system(s) to utilise to construct warnings, it is important to compare systems in their ability to forecast the occurrence and severity of extreme weather events. However, evaluating forecasts for extreme events is known to be a challenging task. This is exacerbated further by the fact that high-impact weather often manifests as a result of several confounding features, a realisation that has led to considerable research on so-called compound weather events. Both univariate and multivariate methods are therefore required to evaluate forecasts for high-impact weather. In this paper, we discuss weighted verification tools, which allow particular outcomes to be emphasised during forecast evaluation. We review and compare different approaches to construct weighted scoring rules, both in a univariate and multivariate setting, and we leverage existing results on weighted scores to introduce weighted probability integral transform (PIT) histograms, allowing forecast calibration to be assessed conditionally on particular outcomes having occurred. To illustrate the practical benefit afforded by these weighted verification tools, they are employed in a case study to evaluate forecasts for extreme heat events issued by the Swiss Federal Office of Meteorology and Climatology (MeteoSwiss).
Clément Chevalier, David Ginsbourger, Olivia Martius
Modeling the joint distribution of extreme weather events in multiple locations is a challenging task with important applications. In this study, we use max-stable models to study extreme daily precipitation events in Switzerland. The non-stationarity of the spatial process at hand involves important challenges, which are often dealt with by using a stationary model in a so-called climate space, with well-chosen covariates. Here, we instead chose to warp the weather stations under study in a latent space of higher dimension using multidimensional scaling (MDS). The advantage of this approach is its improved flexibility to reproduce highly non-stationary phenomena, while keeping a tractable stationary spatial model in the latent space. Two model fitting approaches, which both use MDS, are presented and compared to a classical approach that relies on composite likelihood maximization in a climate space. Results suggest that the proposed methods better reproduce the observed extremal coefficients and their complex spatial dependence.
Jonathan Koh, Daniel Steinfeld, Olivia Martius
When extreme weather events affect large areas, their regional to sub-continental spatial scale is important for their impacts. We propose a novel machine learning (ML) framework that integrates spatial extreme-value theory to model weather extremes and to quantify probabilities associated with the occurrence, intensity, and spatial extent of these events. Our approach employs new loss functions adapted to extreme values, enabling our model to prioritize the tail rather than the bulk of the data distribution. Applied to a case study of Western European summertime heat extremes, we use daily 500-hPa geopotential height fields and local soil moisture as predictors to capture the complex interplay between local and remote physical processes. Our generative model reveals that different facets of heat extremes are influenced by individual circulation features, such as the relative position of upper-level ridges and troughs that are part of a large-scale wave pattern. This enriches our process understanding from a data-driven perspective. Our approach can extrapolate beyond the range of the data to make risk-related probabilistic statements. It applies more generally to other weather extremes and offers an alternative to traditional physical and ML-based techniques that focus less on the extremal aspects of weather data.
Monika Feldmann, Michael Blanc, Killian P. Brennan, Iris Thurnherr, Patricio Velasquez, Olivia Martius, Christoph Schär
Supercell thunderstorms are the most hazardous thunderstorm category and particularly impactful to society. Their monitoring is challenging and often confined to the radar networks of single countries. By exploiting kilometer-scale climate simulations, a first-of-its-kind characterization of supercell occurrence in Europe is derived for the current and a warmer climate. Despite previous notions of supercells being uncommon in Europe, the model shows ~700 supercells per convective season. Occurrence peaks are co-located with complex topography e.g. the Alps. The absolute frequency maximum lies along the southern Alps with minima over the oceans and flat areas. Contrasting a current-climate simulation with a pseudo-global-warming +3$^\circ$C global warming scenario, the future climate simulation shows an average increase of supercell occurrence by 11 %. However, there is a spatial dipole of change with strong increases in supercell frequencies in central and eastern Europe and a decrease in frequency over the Iberian Peninsula and southwestern France.