A Research Network
The overall objective of Module A is to significantly improve the understanding of scale interactions and links between atmospheric circulation, their natural variability and the resulting extreme events. This comprises identifying large-scale atmospheric patterns associated with the extreme weather events and understanding whether these large-scale attributes are consistent with changes in the planetary circulation and with the identified dynamics. This analysis is required for a robust estimate of future changes.
Module C focuses on extreme events that have an impact on socio-economic systems. This means that it is not only a local extreme value of a meteorological parameter that is of interest, but the combination of specific environmental attributes that make it impact relevant. This may include the spatial extent of extremes, nonlinearities in dependencies, the combination of meteorological parameters, so-called compound events, the influence of non-climatic human factors or vulnerability and exposure.
The project SEVERE investigates the physics, processes and scale dependency of very extreme precipitation events. Very extreme precipitation events with very long return periods (e.g. 100 years) can potentially cause large damages, especially when followed by regional or large scale flooding. This is crucial in a warming climate since atmospheric physics shows that warmer air contains more water than colder air (as described in the Clausius-Clapeyron equation). Hence a larger water content in the air masses brings an increased potential for precipitation extremes. However, this effect is not the only factor since the future development depends as well on the large and regional scale evaporation, atmospheric stability conditions
and large-dynamics dynamics. The period, for which reliable observations exist (~ 50 years), is too short to derive robust estimates on longer time-scales. Therefore, SEVERE will use the data from existing large ensembles of regional climate simulations from German and International projects (MiKlip, CMIP-5/6, CORDEX). The project is structured into three phases: i) The characterization of intensity, extension and duration of observed extreme precipitation events over Europe with respect to their temporal and spatial distribution. ii) Evaluation of the potential of climate simulations to reproduce the relevant features of extreme precipitation as well as the large and
regional scale processes. iii) The results will then be applied to the existing large ensembles of climate simulations to identify a sufficient number of very extreme precipitation events.
Heavy, large-scale precipitation events and associated floodings represent one of the greatestnatural hazards for society in Europe. The rarity of the most extreme precipitation events, whichcause particularly severe damage, makes systematic scientific investigations difficult. The processesbehind such extreme precipitation events are therefore not yet fully understood. This project aims toinvestigate the most extreme and relevant precipitation events in European river basins regardingtheir causative processes by using a new, innovative approach to data collection. Subsequently,differences to strong but less extreme events will be highlighted and future changes of these mostextreme precipitation events in a warmer climate will be estimated. Different meteorological parameters will be used to study the large-scale atmospheric circulation and the associated moisture transport during these events. Weather systems such as cyclones, fronts and blocking anticyclones will be objectively identified to classify the circulation anomalies. In order to better understand the differences to less extreme precipitation events, these analyses will also be applied to 1- to 20-year events and compared with the 100-year events. Finally, possible future changes of the most extreme precipitation events in a warmer climate will be estimated under a "business as usual" scenario using climate simulations of the Community Earth System Model.
In the last decades, dramatic Arctic climate change has occurred. Warming of theArctic, much faster than the global average (Arctic amplification) is related tosignificant sea-ice retreat. At the same time, a large number of extreme weather andclimate events occurred in the Northern Hemisphere mid-latitudes. Whether changesin the frequency and magnitude of these extremes is natural or due to anthropogenicinfluence is under debate. Many types of extremes like heat waves or extremeprecipitation are related to anomalies in the large-scale atmospheric circulation.Recent studies emphasize that Arctic amplification can contribute to changes in themid-latitude atmospheric circulation. This project advances our understanding of themechanisms underlying the chain of linkages between Arctic climate change andextremes over Central Europe. It investigates, whether changes in these extremescan be attributed to Arctic climate changes, in particular sea ice loss. We put a focuson the links between sea ice changes, large-scale circulation anomalies, and theirdirect and indirect effects on the occurrence and severity of meteorological extremes.
In PERSEVERE we aim at improving our physical understanding of atmospheric waveresonanceevents in summer. Such events are often associated with persistent summer extremes in Europe, such as heatwaves, droughts and storms that bear important environmental and societal risks. For instance, in 2018, Western Europe experienced a record long hot-dry period lasting essentially the full summer causing massive impacts on agricultural production. There is evidence that weather persistence has indeed increased in Europe over the last decades, associated with changes in atmosphere dynamics, but the uncertainties are large and the exact drivers not well understood. Here, with the use of Machine Learning methods in combination with climate models, we will assess how hemispheric-scale jet regimes are connected to wave-resonance events, we will extract causal precursors, and we will derive emergent constraints in order to reduce uncertainty in future model projections. This project aims at answering the questions: i) How often do double-jet regimes and or/blocking situations coincide with each other and with wave-resonance events? ii) How do such atmospheric configurations affect extreme weather in central Europe? iii) Are there trends in the occurrence of double-jet, blocking, and wave-resonance events? iv) What are the causal drivers (e.g. Arctic Amplification, early-season snowmelt) of waveresonance events in observations and models? v) How well do models simulate such events and can we improve model simulations with emergent constraints derived from causal relationships?
This project deals with the role of atmospheric dynamics in shaping present and future heat waves in Germany. Climate projections show with high confidence that heat waves intensify over the 21st century – a trend that can already be observed nowadays. While local effects of soil moisture deficits have been well established as a contributing factor to future more frequent and more intense heat waves, few studies have been conducted that focus on more remote factors, e.g. via the influence of a changing atmospheric circulation on these events. Necessary conditions for heat waves in Central Europe are slow-moving or stationary high-amplitude Rossby wave trains in the upper-level flow. The upper-level ridges correspond with a strong high-pressure system in the lower troposphere. Therefore, group and phase velocities of Rossby wave packets, the occurrence of Rossby wave breaking and blocking are diagnosed in reanalyses and higher-resolution climate projections (CMIP5, CMIP6 HighResMip, MPI-ESM, CESM-Large Ensemble, CORDEX II Europe). In addition, the origin, transport and modification of air masses and air parcels are quantified by trajectory analyses with the program LAGRANTO. Thereby, the important physical processes for the evolution and persistence of heat waves can be compared with each other and uncertainties in the climate projections can be quantified and potentially reduced. Finally, the question whether atmospheric dynamics intensify the most extreme heat waves even more than expected, can be answered.
Intense mid-latitude cyclones are one of the main weather hazards in Europe. They areassociated with strong winds and heavy precipitation which can lead to wind damage andflooding. It is therefore important to understand how mid-latitude storms will respond toclimate change, in order to predict future weather risks and to guide climate changeadaptation strategies. This project investigates cyclone-scale features like strong winds and fronts, with a focus on windstorms in the Atlantic-European sector. Overall, CyclEx consists of two parts. In the first part extreme cyclones are analyzed in global climate models (GCMs), in the second part they are analyzed in idealized simulations. These two parts enable us to bridge the gap between low-resolution climate models and small-scale processes that are better resolved in high-resolution simulation. In general, we aim to better understand (1) the cyclone dynamics under global climate change, (2) the relative impact of environmental and diabatic processes on the cyclone scale and (3) the consistency, ordifferences, of cyclone changes in high- versus low-resolution simulations.
The aim of this project is to link severe convective storms (SCS) in Europe with large-scale atmospheric processes and to investigate the influence of these processes on the annual variability of SCS both in the past and in the future. Furthermore, this project investigates how large-scale mechanisms determine clustersof SCS on scales of several days to weeks (referred to as serial clustering). Knowledge on thesemechanisms is still insufficient, but a prerequisite for estimating robust statements about long-termchanges and trends in the SCS frequency. VarCluST intend to answer two major scientific questions: i) which large-scale processes (e.g., teleconnection patterns or certain weather types such asatmospheric blocking) have an influence on the high observed annual variability of SCS in the pastand which changes in thunderstorm frequency and relevant processes can be expected in thefuture? ii) which atmospheric mechanisms influence the increased occurrence of SCS within a few daysup to several weeks (serial clustering)? The obtained knowledge will be transferred to an ensemble of regional climate models to estimate the probability and persistence of SCS and SCS-related large-scale processes in future decades.
In recent decades, Europe has experienced an increasing number of extremely warm summers. This tendency has largely been attributed to the anthropogenic increase of greenhouse gas emissions, and is expected to be accentuated as global warming continues to increase. The changes in the frequency and amplitude of the European heat extremes depend not only on the level ofglobal warming, but also on the large sub to multi-decadal variability of the climatesystem. In this project we will quantify the contribution from sub to multidecadalvariability to the frequency and intensityof European heat waves and will investigate their drivingmechanisms. For this purpose we will use observations and simulations from the Max Planck Institute Grand Ensemble (MPI-GE). By evaluating the large samples of extremes under different climate conditions simulated with the MPI-GE, we can identify how the decadal variability in the climate system, such as the North Atlantic Ocean heat transport variability and respective recurring circulation patterns, affect the extreme summer temperatures and their driving mechanisms: the dynamical large scale atmospheric state and the local thermodynamical effect of moisture limitation. Ultimately, our goal is to establish attribution and projection frameworks that account for not only mean global warming, but also for the sub to multi-decadal variability in the climate system.
This project aims at an understanding of regional impacts of global climate change on extreme sealevel heights in the North Sea. Climate change effects in the atmosphere and ocean may increase therisk for local flooding due to storm surges or high water levels in rivers and estuaries. Based on modelresults of high-resolution climate projections for the 21st century, both driving mechanisms andleading variability modes of sea level extremes at the continental North Sea coast shall be identified,including an in-depth analysis of potential future changes in the occurrence and dynamics ofcorresponding weather conditions. Particular focus will be laid on the investigation of the interplay ofexternal and internal storm surges, tides and hydro-meteorological events, which are consistentlysimulated by a regionally coupled atmosphere-ocean climate system model. Large-scale driving mechanisms of extreme sea levels will be identified and evaluated by incorporating a hindcast simulation forced with atmospheric reanalysis data. Projected climate change signals due to anthropogenic greenhouse gas emissions as well as the natural variability of the dynamic system will be assessed over a wide range of spatial and temporal scales (regional to large-scale, inter-annual to multi-decadal). Estimates of potential future changes based on such a comprehensive approach with respect to physical process representation and statistical validity significantly improve the current understanding of climatic drivers of extreme sea levels.
The project aims to build and continuously expand an integrated database on damaging weatherconditions on the basis of past and recent observational datasets (reanalysis data) as well as recent and future climate model projections. This starts from well-established extreme indices and integrating novel extreme indices through collaboration with the other projects from Module C focusing on impacts of multiple hazards considered in Module C. The project COO coordinates Module C and will integrate individual Module C work package results to build the database for damaging weather conditions in central Europe. As a key scientific question, the project will address clustering of damaging weather events which are particularly relevant in the (re)insurance context since clustering of damaging events can have severe economic implications.
Flash floods are among the most destructive and ubiquitous natural hazards in Central Europe. Mostly caused by extreme convective rainfall, we still lack a systematic understanding of how their impacts (in terms of damage to buildings and infrastructure) are related to the spatiotemporal attributes of the triggering rainfall (duration, spatial extent, intensity, and location), and the local terrain conditions. To address that gap, CARLOFFF aims at extracting a comprehensivecatalogue of convective events and their spatiotemporal attributes from almost 20 years of DWD weather radar data, and at linking these attributes to an extensive database of reported impacts from various sources. Furthermore, we attempt to map the regional occurrence of such impact-relevant events to large-scale atmospheric conditions (as represented by climate reanalyses). On that basis, it might be possible to detect past and future changes in both frequency and intensity of impact-relevant atmospheric conditions.
The project delivers a comprehensive comparison of different methodologies to model climate and crops at the subnational level in Germany at various temporal scales. It proposes an innovative integrated approach to better understand and characterize the impacts of compound extreme events. CROP generates new high-resolution data sets, based on best-performing hybrid approaches to bias-correct model simulations and downscale them to relevant spatio-temporal scales for the assessment of extreme weather impacts on crop yield in terms of variability and losses. A better understanding of how compound extremes influence crop production sets the basis for the development of an integrated seasonal crop yield forecasting system, an important and accurate tool to inform end-users. The results will be of relevance for climate change impact assessments, to guide extension services and for policy makers to reduce market volatility, avoid price spikes and support crop breeding efforts.
There is great uncertainty not only with regard to changes in extreme climate and weather conditions per se, but also with how extreme effects taking place simultaneously interact in different sectors and, eventually, reinforce one another. EXIMO will focus on the relationships between climate and weather extremes and their extreme effects, which may pose a significant risk to social resilience to climate change. The research work in EXIMO will be closely coordinated with the module C consortium and in particular the sub-project CROP. EXIMO aims to clarify the importance of interactions between different areas of climate impact research in the context of climate and weather extremes, since the mostly separately considered components of the land biosphere and the water balance actually interact. For example, agricultural water withdrawals are particularly important for maintaining agricultural production and resilience, especially in periods of drought, but at the same time water availability may be particularly low, so that extreme situations can arise due to the interaction. The aim is to identify such events and their meteorological framework conditions, to quantify the interactions using existing data sets for simulated and observed extreme events as well as targeted new simulations and to improve process understanding.
Extreme, large-scale river floods typically affect more than one river basin. Although such trans-basin floods are highly relevant for national disaster risk reduction and insurance, the knowledge about the processes triggering these extremes is very limited. The project aims to quantitatively understand the generation and impact of trans-basin floods, and how they are linked to climate change and variability. It will analyse large-scale floods in Central Europe for the observational period 1950-today and future periods (2030-2060, 2070-2100) under climate change. A model cascade, representing the processes from the large-scale atmospheric situations through the catchment and river processes to the damage, will provide ensembles of trans-basin floods, based on changes in the frequency and persistence of weather patterns and pattern internal trends in temperature and precipitation. The project will pinpoint major changes in weather conditions resulting in distribution changes of flood characteristics and related impact.
Wind throws/breaks and drought effects are major climate risks for forests in Central Europe (and thus also in Germany) in the 21st century. In a scenario of increased future cyclonic activity, decreased return periods of severe windstorms, higher variability in extreme wet-dry soil moisture conditions and associated compound events result in potentially severe ecological and economic damage. These include reduced Ecosystem Services (ESS) like water retention, erosion protection, or input of high-quality water to drinking water reservoirs. In the proposed study, a combination of process modeling (soil water status with BROOK90), planetary boundary layer modeling (HIRVAC) and diagnostic wind modeling (WITRAK) will be applied to zoom into areas representing various climate regions of Germany subject to well- documented storm damage since 1990. Combined non-linear statistical approaches will be used to establish relationships between wind throws/breaks and the prevailing preconditions (soil moisture regime, local wind field, etc.) while taking stand characteristics (age, structure, species composition, etc.) into account.
The project analyzes existing simulations from European regional climate model downscaling experiments (CORDEX) plus self-provided new climate simulations, incorporating direct water budget modifications at the land surface and in the subsurface, in particular, human water use (including urban) related to groundwater pumping and irrigation. The analyses are accompanied by additional, event-based simulations, applying a (non-identical) twin model setup of varying complexity to further extract long- and short- range relations in space and time between subsurface hydrodynamics and land surface processes with the atmospheric circulation under direct human modification. The project will also analyze the high- resolution Hans-Ertel-Centre reanalysis data (COSMO-REA6). Software and analyses will be made publicly available via Module D in ClimXtreme.
Landslide processes in Central Europe are associated with high damage on road, railway and building infrastructure as well as casualties. While the general susceptibility to landslides is determined by geological and geophysical conditions, meteorological factors frequently determine the triggering of the hazard. High soil moisture preconditions, caused by rapid snow melt or abundant precipitation, followed by high intensity rainfall have been identified as important triggers. The project will determine regional thresholds for combinations of meteorological factors controlling landslide frequency in Germany using a soil water balance and slope stability model as well as statistical approaches. The basis for all analyses are landslide records of the German landslide database, which compiles damage on infrastructure (road, railway, buildings) and casualties as well as controlling factors and triggers. The results will be used to systematically deduce occurrence probabilities of relevant meteorological factors under present day and climate change conditions using an ensemble of regional climate simulations and to estimate the impacts on landslide frequency and magnitude.
Winter windstorms are among the most dangerous and costly natural hazards in Central Europe. Compared to convective events, the large extent of affected areas makes them particularly relevant in terms of the risks from both an insurance and an economic point of view. The idea of this project is to improve the analysis of weather patterns and -sequences leading to wind-induced damages using a common approach for three different types of storm impacts, namely building damages, forestry damages and railway disruption risks as an example for (secondary) damages due to wind throw.
This project concentrates upon two major lines of tasks. First, the coordination of ClimXtreme in concert with the coordinators of Module A, C and D together with the coordination of the sixteen subprojects within Module B. Scientifically, B1.1 will contribute to the advanced methods of detection and attribution of climate change by anthropogenic influences using a Bayes statistical approach. The methods will include a single event attribution based on an explicit Bayesian likelihood modelling for observed heatwave and wind storm cases given the factual and counterfactual scenario. The data basis will be given by the MiKliP decadal prediction system which can be combined into a 75 member lagged ensemble for about 50 years. Additionally, a set of simulation data will come from the long term (1880 – 2010) historical full forcing, historical natural forcing and historical anthropogenic CMIP5 simulations besides the preindustrial control run to compare different factual scenario (historic full forcing, historic anthropogenic) with different counterfactual (preindustrial control, historic natural) simulations in the Bayesian sense. Observations will be taken from station data and the regional reanalyses REA6.
Project Website: Website
Participating Institutions: Institute of Geosciences, University of Bonn
Contact: Dr. Ieda Pscheidt
of weather and climate. The types of events considered are heat waves, heavy rainfall events and droughts. A major goal is to evaluate the applicability of attribution studies to the output of climate models on a range of spatial resolutions (GCMs, RCMs) and time scales (climate model projections, decadal and seasonal forecasts). Finally, a concept for the operationalization of the attribution of weather and climate extremes is developed. The overall objective is to carry out essential preliminary work for this purpose and to develop a prototype system. For this purpose, existing components are tested for their applicability, adapted where possible and newly developed where necessary. It is essential to develop a process chain that essentially works without the intervention of an operator. A focus is also placed on the analysis of the robustness of the system for a large number of data sets and under different climatic conditions.The project aims at establishing a direct collaboration with the group worldweatherattribution.org and the activity "Copernicus Proof of Concept study for an Extreme Events and Attribution Service".
Climate change may impact societies in particular by extreme events; extreme precipitation andlightning from deep convection (thunderstorms) are a particular threat. Anthropogenic modification ofthe atmospheric composition may drive changes in these events. Different from greenhouse gases andglobal warming, aerosol particle emissions may directly impact deep convection. This project will (i)assess the impact of aerosols on deep convection as simulated by the ICON regional model withexisting large-domain large-eddy simulations, a multi-model ensemble, and ground-basedobservations as reference, and (ii) investigate in a probabilistic attribution framework (ensemblemodelling of factual and counterfactual conditions) the impact of aerosols on extreme rain andlightning. Possible work in a second phase will investigate these in a long-term and global framework.
This project detects and analyses heavy precipitation events (HPE), both single convective events and intense precipitation episodes, with the potential to small- and large-scale flooding over Central Europe. It especially focuses on the analysis of related physical processes relevant for the formation and development of these HPEs, in particular the impact of land-atmosphere (L-A) feedback. The main objective of this project is to analyze the ability of the relatively new ReKliEs-De ensemble of regional climate simulations to reproduce intensities, frequencies, and durations of heavy precipitation events. For a historical climatological period, the classical evaluation strategy is extended to L-A metrics, which characterize the influence of local to regional processes as well as the role of soil, land cover, and the atmosphere leading to heavy precipitation. Furthermore, changes in the statistical and physical quantities of these events are investigated and related to modifications of the forcing mechanisms and L-A feedback for two different climate change scenarios. We will investigate the influence of an improved representation of L-A feedback mechanisms and extreme precipitation events in a series of convection-permitting (CP) simulations on the kilometer scale where a better representation of orography, soil and land-cover heterogeneities, temporal and spatial evolutions of the surface layer and the planetary boundary layer (PBL), and thus, convergence zones and convection initiation can be expected. Our research will lead to a deeper understanding of intensive and extreme precipitation and thus also provide new insights into climate change.
In the past two decades, Europe has suffered from major heatwaves: the extreme heat event that occurred in 2003 caused around 30.000 deaths, and in 2015, 2018 and 2019 many European cities were subject to record-breaking temperatures. Hence, being able to know in advance that a heatwave might occur would be of tremendous benefit for planning. However, is it possible to make informed statements about a potentially upcoming heatwave several months in advance? One potential candidate for memory in the climate system that might enable us to make statement about a potentially upcoming heatwave, is the Atlantic Ocean. The present project aims to tackle a potential influence of the North Atlantic ocean on European summer extremes in a systematic way. As a first step, we will analyze the Max Planck Institute for Meteorology Grand Ensemble (MPI-GE), consisting of 100 simulations from 1850 to 2005. We will compare the frequency and occurrence of heat extremes to re-analysis. As a second step, we will implement statistical methods to automatically detect past heatwave events, and analyze the Atlantic Ocean’s behavior during spring. Using state-of-art methods, such as Machine Learning (ML) and non-gaussian verification techniques, we will investigate which regions of the Atlantic Ocean play a major role in the occurrence of heatwaves over Europe. With such a robust identification and description of potential North Atlantic SST pre-cursors for European summer extremes, we will aim to improve seasonal heatwave predictions in a final step.
Convective hazards such as large hail, severe wind gusts, tornadoes and heavy rainfall are responsible for high economic damages, fatalities and injuries across the world, in Europe, and in Germany. There are insufficient observations to determine whether trends in such local phenomena exist, but recent studies suggest that conditions associated with such hazards have become more frequent across large parts of Europe in recent decades. These conclusions are in part based on work with Additive Regression Convective Hazard Models (AR-CHaMo) that have been developed using state-of-the-art reanalysis data and observations collected in the European Severe Weather Database (ESWD). The CHECC project improves AR-CHaMo by using newer reanalysis datasets with higher spatial and temporal resolutions, such as ERA5, COSMO-REA6 and MERRA2. CHECC uses the models to investigate if significant trends in modelled hazard occurrence can be detected both in the past and in future climate projections. Furthermore, CHECC studies which part of these trends can be attributable to changes in tropospheric flow patterns, by assessing the impacts of any detected changes on the underlying physical drivers of convective events. Finally, CHECC will explore the use of convection-permitting reanalysis data, such as COSMO-REA2. This is of particular interest as climate projections are gradually becoming available at convection-permitting module resolutions.
The frequency and intensity of extreme precipitation are critical factors for the assessment of future impacts due to rainfall extremes. Other event characteristics can also play an important role: for example, the duration, spatial extent and areal precipitation volume of the event. To fully understand how extreme precipitation may change in a future climate, an analysis encompassing all the properties of extreme precipitation is thus necessary. A two-pronged approach adopting both classical and Lagrangian perspectives on extreme precipitation is thus necessary. In the Lagrangian perspective, precipitation events are viewed as features in space-time to be identified and tracked via a tracking algorithm while assessing their characteristics relative to the transiting storm centre. Intense precipitation events can thus be characterized in terms of their spatial extent, total precipitation volume, mean and maximum intensity, occurrence rate, and – if tracked – also lifetime, distance travelled, speed and area covered. The overarching aim is to investigate and understand potential changes of these feature-based characteristics within a changing climate. For convective precipitation, many of the important underlying processes cannot be directly simulated in standard climate models. Our analyses are thus based on climate simulations, both historical and projections, which have been dynamically downscaled by convection-permitting models to convection-permitting resolution, as well as convection-permitting resolution reanalyses.
Precipitation events can have very different characteristics, from long-lasting light drizzle to short but intense precipitation. For every precipitation event duration, extremes can be defined using extreme value statistics, resulting in intensity-duration-frequency (IDF) relationships. A popular approach to IDF curves is the estimation of extreme value models for individual durations separately and subsequently and, in a second separate step, modelling the duration dependence for fixed quantiles (or return-levels) individually. This leads necessarily to inconsistencies such as a crossing of IDF curves associated to different quantiles (return levels): i.e. a 0.9-quantile being larger than a 0.95-quantile. We use a single consistent model based on a duration-dependent Generalized Extreme Value Distribution (d-GEV) to estimate the IDF relationships for all durations simultaneously (Koutsoyiannis, 1998) and compare different approaches to model the duration dependency of the GEV. W statistical e aim for a meaningful estimation of the IDF relationships across all relevant event durations and for ungauged sites through spatial covariates in a spatial model. To further investigate the parameter range, a Bayesian Hierarchical Model (BHM) will be set up, including a Gaussian Process for spatial modelling. Conditional on large-scale atmospheric flow indices, IDF relationships in a changing climate shall be obtained locally (i.e. at the gauge level) and results will be presented on an interactive web-interface for end-user’s convenience.
The aim of SAVE is to determine the role of atmospheric characteristics, structures and processes in the generation, amplification or attenuation of extreme events by applying a multivariate approach. Especially with regard to the spatial resolution of climate simulations, it is to be expected that the climatological representation of extremes is suboptimal. Therefore, the project goals are: i) determination of scale-dependent atmospheric drivers for extreme events by examining the relationship between atmospheric conditions and processes and the occurrence of extremes on different spatio-temporal scales, ii) analysis of past changes in the occurrence of extremes with respect to changes in the identified scale-dependent drivers, and ii) transfer of the identified relationships of drivers and extremes to climate projections in order to enhance the prediction of future changes in the frequency and intensity of extreme events. Our analysis is based on a multi-scale approach using data sets with different spatial resolutions. This approach allows for the estimation of drivers of extremes on large (global) scales using global reanalysis, continental scales using regional reanalysis and local scales using convection-resolving reanalysis. Furthermore, the planned approach exploits the spatio-temporal drivers identified using reanalysis data (as approximate for the true state) to enhance (regional) climate predictions by applying or transferring the determined statistical relationships between drivers and extremes.
Precipitation extremes are a space-time phenomenon, but have been statistically treated in the past mainly from single point observations. Here, we aim at the statistical analyses of the dynamics of space-time extremes and at the development of new statistical approaches to describe it. As one main outcome, area-intensity-duration-frequency (AIDF) curves will be estimated for past, present and future times. In addition, questions regarding the temporal development of theses extremes, clustering phenomena, unusual events will be addressed. The research utilizes especially high space-time resolution precipitation data from weather radar in addition to the rain gauge and climate model re-analyses data.
Websites: 1, 2
Detection and attribution (D&A) of climate change requires a compact description of the spatio-temporal climate state. While, in high dimensional spaces, particularly on small scales, internal climate variability is typically too large for a climate change signal to be significant, a suitable reduction of degrees of freedom can improve the signal-to-noise ratio and, thus, increase the potential to detect less strong signals in this setting. Therefore, in our project, we analyze and develop methods for information compression of high-resolution spatial weather variables and integrate them into the D&A framework for extreme weather related to moist deep convection. So far, most of the available strategies aim at the description of the bulk of the distribution while our analysis focuses on the tail behavior. Thus, we compare different methods and assess their assets and drawbacks with respect to extreme weather. The methods investigated comprise data adaptive decomposition such as principal component analysis, filter approaches using wavelet decomposition, or dynamical decomposition based on reduced dynamical models. Furthermore, we consider various spatial extreme value models. In particular, we develop new statistical models describing the spatial dependence structure of single extreme events based on Pareto processes.
In this subproject, we develop a methodological framework rooted in complex system science and machine learning to study spatiotemporal patterns of multi-variate and compound extreme events. The framework will be applied to observational and reanalysis data as well as climate model simulations to (1) investigate teleconnections in terms of spatial synchronization patterns of heavy precipitation and storm events and their dependence on large-scale ciruculation patterns with a focus on the predictability of these events; (2) assess the representation of (compound) precipitation and storm extremes and their teleconnections in state-of-the-art general circulation models via comparisons with corresponding observational data; (3) estimate the stability of such teleconnection patterns over time both in historical observations and reanalysis, as well as in climate model projections, to identify potential changes caused by anthropogenic climate change.
The statistical analysis of extreme weather events often ignores a proper accounting for dependencies and heterogeneities induced by the spatio-temporal nature of respective data sets. For instance, taking temporal dependence into account is important for the statistical assessment of the inter-arrival times between extreme events. Furthermore, statistical methods for data collected at various locations and times can be improved substantially if information is available about which locations share similar characteristics and which model parameters can be assumed to be constant over space or time. This project is divided into two parts. The first part, "Statistical inference for serial clustering of intense storms and heavy precipitation“, investigates a generalized extreme value theory model, so-called Continuous Time Random Maxima (CTRM), which allows for heavy-tailed inter-arrival times so that clustering of extreme events occurring in bursts can be captured. The second part, "Homogeneity analysis for spatio-temporal weather extremes", provides statistical tools to detect and make use of homogeneity information, in particular by investigating estimation strategies based on regularization techniques.
Websites: 1, 2, 3
In climate science, compound events can be i) two or more extreme events occurring simultaneously or successively, ii) combinations of extreme events with underlying conditions that amplify the impact of the events, or iii) combinations of events that are not themselves extremes but lead to an extreme event or impact when combined (Seneviratne et al. 2012). Here, we focus on the identification and statistical modelling of compound events resulting from i) heavy precipitation with windstorms, ii) heavy precipitation with high river discharges, iii) heavy precipitation with storm surges, and iv) precipitation-induced high river discharges with storm surges. The four combinations are drivers for major societal risks from hydrological natural disasters. Thus, the project aim is to gain profound knowledge of past and future changes of the probabilities and magnitude of such compound events. This is of special importance, e.g. for building resilient societies. The project ProComE is organized into three work packages: i) Data compilation and basic analysis, ii) Univariate extreme value methods, and iii) Analysing probabilities of compound events. The results of the project will improve the understanding of past and future changes of probabilities of compound events. Since natural disasters, triggered by hydrological compound events, are a major source for today’s societal risks, the results of this project will help to reduce these risks and hence give support to establish more resilient societies.
Standard statistical methods in climatology have certain weaknesses in detecting and analyzing present and future climate extremes. The goal of MarNet is therefore to reduce these deficiencies and complement conventional approaches by applying new and innovative methods, which are not commonly used in the context of extreme events, namely Climate Networks and Markov Chains. A Climate Network consists of nodes which are connected by edges. Nodes can either be observation sites or grid points of a climate model, and connections between such nodes are established if time series of meteorological data are there correlated. In this way, statistical similarities between nodes can be revealed, by which physical interrelations in the climate system can be analyzed. The method is therefore particularly suited to identify spatial (hot spots; changing locations) and temporal patterns in present and future extreme events. A Markov Chain is a time series of different states of a system (e.g. extreme or not-extreme) in which the probability of a system to change its state is incorporated in a transition matrix. By using different descriptors of a Markov Chain, certain characteristics of a system can be determined, such as the tendency of a system to stay in a given state, or to return to this state, and the predictability of a state transition. In this way, Markov Chains can be used to analyze the frequency, duration and regularity of present extreme events and describe their future change.
Extreme weather and climate events are of great general societal interest since they cause significant economic damages and fatalities each year. The description and prediction of extremes is involved since these events can be correlated. European windstorms, for example, occur more often in bunches than if they were independent events (Franzke 2013, Blender et al. 2015). The observed clustering of extremes is caused by meteorological processes like Rossby waves and downstream development. Blender et al. (2015) have shown that the return times between severe windstorms are non-exponential and have to be modeled by a fractional Poisson process instead of the standard Poisson model for rare events. Extreme events are typically examined from a Eulerian perspective by determining the extreme value characteristics at particular stations or grid points. Here we propose a different and novel approach by taking a Lagrangian view on extremes to grasp their accumulated impact. The vortex tracking (Blender et al. 1997, Sienz et al. 2010) is combined with an analysis of the extreme value characteristics and the clustering properties. The return times between two extreme events are scientifically relevant and important for stakeholders and the insurance industry. We aim to understand and predict severe weather and climate events by combining the Lagrangian cyclone view with newly developed concepts for the serial clustering of extreme events. We will elucidate the driving mechanisms of the serial clustering of cyclones.
To support the scientific activities in ClimXtreme, module D takes over the coordination of software and data management. The priority of the data-related work in CoDaX is to support the ClimXtreme partners in the evaluation and application of new data products by providing and integrating them into a central evaluation system. One highly relevant data source are ground-based meteorological observations for analyzing the extremes of the last decades, especially when looking at centennial time scales. Second, for the assessment of regional extremes in Europe and Germany regional reanalyses have come up as opportunity for a detailed analysis of extreme events. Besides, several gridded data sets as well as global products are considered and will be made available to all ClimXtreme members. In addition to provision and integration, the development of methods for the evaluation and assessment of the data sets with regard to extreme events will be part of CoDax as well. CoDaX will support ClimXtreme by identifying relevant definitions of extreme events and testing the suitability of the station data and reanalysis data for the analysis of extremes based on these definitions.
The ClimXtreme project (climxtreme.de) has been conceived with the aim of generating climate knowledge to contribute with an improved assessment of damage-related extreme weather events in Central Europe: it will examine events that have taken place in the near past to assess the changes in both frequency and intensity that are possible in the future. These events will be analysed under three different perspectives: the physical processes that cause them (module A), their statistical aspects (module B) and their impact (module C). A fourth module (module D), will support the scientific activities of the other three modules providing a coordinated strategy for modelling, software and data management. In this context, CoSoX supports with software related topics, developing a common and central evaluation system for climate extremes (XCES) allowing for a flexible incorporation of verification routines for modules A,B,C. XCES, hosted at the High Performance Computer of the DKRZ, provides the infrastructure to connect the scientists through its common web and shell platforms allowing them to browse and produce datasets, plugins and results within the platform. CoSoX at the DKRZ in cooperation with the FU Berlin will develop and provide for templates and basic routines for software development and data standardization, (potentially) adapting and providing already existing routines as well. In conjunction with CoDaX, the other project in module D, CoSoX also works on the integration and provision of datasets needed for the project.