4-9 September 2022, Bonn, Germany
Enter gather.town


Climate change detection, assessment of trends, variability and extremes

Society will feel the impacts of climate change mainly through extreme weather and climate events, such as heat waves and droughts, heavy rainfall and associated flooding, and extreme winds. Determining from the observational record whether there have been significant changes in the frequency, amplitude and persistence of extreme events poses considerable challenges. Changes in the distributional tails of climate variables may not necessarily be coherent with the changes in their mean values. Also, attributing any such changes to natural or anthropogenic drivers is a challenge.

The aim of this session will be studies that bridge the spatial scales and reach the timescales of extreme events that impact all our lives. Papers are solicited on advancing the understanding of causes of observed changes in mean climate, in its variability and in the frequency and intensity of extreme events. In particular, papers are invited on trends in the regional climate of Europe, not just the mean, but variability and extremes, often for the latter measured through well-chosen indices.

Convener: Martine Rebetez | Co-conveners: Albert M.G. Klein Tank, Monika Lakatos
| Wed, 07 Sep, 09:00–10:30 (CEST), 11:00–13:00 (CEST)|Room HS 3-4
| Attendance Wed, 07 Sep, 14:00–15:30 (CEST) | Display Wed, 07 Sep, 08:00–18:00|b-IT poster area

Wed, 7 Sep, 14:00–15:30

Chairpersons: Albert M.G. Klein Tank, Martine Rebetez, Monika Lakatos

Onsite presentation
Seánie Griffin and Keith Lambkin

A key goal in the developing area of Climate Services at Met Éireann is the delivery of relevant climate information that can assist policy decision making in Ireland.

One such area is the provision of indices associated with climate extremes. Here the ETCCDI indices have been used to assess trends in extreme temperature and rainfall events at locations around Ireland, based on daily observations.  The primary aim is to provide easily accessible information for end users to assess the potential vulnerabilities/opportunities posed to their sectors by Ireland’s changing climate; be it in the form of longer growing seasons, less frosts or greater heat stress.

Reductions in cold extremes are found to be generally more significant than increases in warm extremes, with the warming of minimum temperature extremes being slightly more robust than for maximum temperatures. Warming trends pervade all the temperature indices and the results are generally consistent across the stations analysed.  This is not the case for rainfall indices, with large year-to-year variability and differences between stations. The majority of stations indicate increasing trends in total precipitation, though only a minority of these are found to be statistically significant. There is a large degree of uncertainty in the indices associated with extreme heavy rainfall events.

Specific drought monitoring products have also been developed for users in the water sector, in the form of maps of the Standardized Precipitation Index (SPI) and Standardized Precipitation-Evapotranspiration Index (SPEI) using gridded observational datasets. While the Irish climate is typically associated with abundant amounts of rainfall, two recent drought events (July 2018 and May 2020) caused significant disruption for both agriculture and general water infrastructure. This has focused attention on providing relevant up-to-date information to relevant stakeholders, and resulted in the production of a near real-time operational product of catchment average SPI.

An outline of the framework for producing and disseminating these indices will be presented, along with the main findings of dominant trends in climate extremes for Ireland.

How to cite: Griffin, S. and Lambkin, K.: Climate Monitoring Indices for Ireland, EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-512, https://doi.org/10.5194/ems2022-512, 2022.

Onsite presentation
Pavel Faško and Ladislav Markovič

Analysis of annual precipitation totals in Slovakia in the last decades of the 20th century and during the 21st century shows a certain, although not very significant, upward trend. Paradoxically, the incidence and severity of the periods with lack of precipitation is also increasing. In this context, it is appropriate to analyze precipitation events resulting in heavy precipitation, whereas one of the consequences of their occurrence is an increase in high daily, resp. multi-day precipitation totals. This paper presents analysis of new record values of the maximum daily, resp. multi-day precipitation totals in Slovakia recorded in the period 2001-2021 with focus on the comparison with previous record values from the period 1951-2000 on more than 450 precipitation stations. The analysis was performed on annual and seasonal time series ​​of maximum daily and multi-day total precipitation totals. The paper presents results of the analysis for those months in which the changes were most significant. Remarkable is, in particular, the increase in the values ​​of the analyzed characteristics of precipitation during the whole winter season (XII – II). Similar development was also detected in other seasons, but was limited to individual months only. The increase in the winter precipitation totals is associated with an increase in air temperature. Warmer atmosphere results not only in the proportional change in the total amount of the solid, liquid and mixed precipitation but modifies character of the winter precipitation. Changes in in the maximum daily and multi-day precipitation totals were more pronounced in the more continental climate of eastern Slovakia.

How to cite: Faško, P. and Markovič, L.: Changes in the maximum daily and multi-day precipitation totals in Slovakia in the current course of the 21st century., EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-180, https://doi.org/10.5194/ems2022-180, 2022.

Online presentation
Santiago Gaztelumendi et al.

UrbanKlima2050 is the most ambitious initiative led by the Basque Country which aims at ensuring the resilience of the territory from a multi-level governance, multi-stakeholder, and on-the-ground climate action approach. This largescale Life Integrated Project was launched at the end of 2019 thanks to a partnership of 20 organisations and an investment of EUR19,8 million. The project partners have set an ambitious target of delivering different actions during 6 years, impacting over 2 million people. The main aim of UrbanKlima2050 project is to contribute to the full implementation of the Basque Climate Change Strategy 2050, developing a low-carbon and climate change resilient region by 2050.

Project actions are grouped in five major blocks: (1) Analyse: to review the Basque KLIMA 2050 strategy through follow-ups and evaluations, with a continuous improvement approach; (2) Define: how, where and when to act to reduce GHG emissions, increase absorption, and achieve territorial resilience; (3) Act: to launch pilot projects at three levels of intervention: coast, river basins and urban/peri-urban areas, scalable to other areas of the Basque Country and to other regions; (4) Empower: to promote climate awareness among the governmental institutions and the community and move them to action; (5) Manage: to create structures to facilitate climate governance and climate change observation and monitoring and define new models for climate governance and launch the Hub for observing and monitoring of climate change in the Basque Country.

Tecnalia is participating on twelve actions, leading five. Here we focus on those aspects related to the Climate change hub action, led by IHOBE, where different activities are developed by Tecnalia, AZTI, Neiker, and other institutional agents in order to establishing a Hub for observation and monitoring climate change at Basque Country level. Particularly, the weather and climate area of Tecnalia works in different task in the atmospheric domain.

Among other topics we cover the identification of the main sources of instrumental information available in relation to the atmospheric characterisation of the territory, including both those for which longer historical records are available and those of a more novel nature that may be useful in the future in an operational perspective. We also design and implement strategies for signals of change and trends detection of the most representative indices and indicators at different level of spatial and temporal aggregation.

In this paper, we share our experience in the development of this action, summarising the key aspects and providing some conclusions in relation to the definition and implementation of the monitoring system for atmospheric-based indicators for the Basque Country. This system will be an essential part of the Basque HUB on climate change to be developed in UrbanKlima2050 project that will end in 2025.

How to cite: Gaztelumendi, S., Gomez de Segura, J. D., Hernandez, R., Martija-DIez, M., and Aranda, J. A.: Climate change monitoring and atmospheric indices in Basque Country: experiences from URBANKLIMA2050 LIFE project., EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-596, https://doi.org/10.5194/ems2022-596, 2022.

Presentation form not yet defined
Classification of extreme temperature events in the presence of seasonality and climate change 
Stefanie Gubler et al.
Onsite presentation
Pavel Zahradníček et al.

Thirty-year periods are treated in climatology as spans with relatively representative and stable climatic patterns, which can be used for calculating climate normals. Annual and seasonal series of circulation types were used to compare two 30-year sub-periods, 1961–1990 and 1991–2020, the second one being strongly influenced by recent global warming. This analysis was conducted according to the objective classification of circulation types and the climatic characteristics of sunshine duration, temperature, humidity, precipitation, and wind speed as calculated for the territory of the Czech Republic during the 1961–2020 period. For both sub-periods, their statistical characteristics were calculated, and the statistical significance of differences between them was evaluated. There was a statistically significant increase in the annual frequencies of anticyclonic circulation types and a significant decrease in cyclonic circulation types during 1991–2020 compared with 1961–1990. Generally, in both 30-year periods, significant differences in means, variability, characteristics of distribution, density functions, and linear trends appear for all climatic variables analysed except precipitation. This indicates that the recent 30-year “normal” period of 1991–2020, known to be influenced more by recent climate change, is by its climatic characteristics unrepresentative of the stable climatic patterns of previous 30-year periods. Annual and seasonal sunshine duration series do no express significant changes between two 30-year normal periods in terms of their variability, characteristics of distribution, density functions and linear trends. Only increase in annual, MAM and JJA means in 1991–2020 compared to the preceding period was statistically significant. Mean, maximum and minimum temperatures display quite different patterns in two 30-year normal periods in accord with recent warming. They are reflected in statistically significant differences in means, characteristics of distribution, density functions and in significant linear trends in 1991–2020 (annual, JJA, SON). It is pronounced particularly for JJA series. Statistically significant decreases in means of relative humidity between two 30-year normal periods (except SON) are reflected in a significant shift of density functions to lower values. But decreasing linear trends were significant only for MAM. Increasing variability in relative humidity was significant for annual and MAM series. Precipitation totals of the both 30-year periods are represented well by insignificant linear trends, no substantial changes in mean and variability as well as in the character of their distribution represented by the density functions. Wind speeds in two 30-year normal periods represents quite different patterns expressed by statistically significant decreasing linear trends in 1991–2020, significant differences in means, partly in variability and skewness (annual, MAM) and in density functions.

How to cite: Zahradníček, P., Brázdil, R., Dobrovolný, P., Řehoř, J., Lhotka, O., Trnka, M., and Štěpánek, P.: Circulation and Climate Variability in the Czech Republic between 1961 and 2020: A Comparison of Changes for Two “Normal” Periods, EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-491, https://doi.org/10.5194/ems2022-491, 2022.

Onsite presentation
Tomas Krauskopf and Radan Huth

While long-term changes in measures of central tendency of climate elements, that is, mean temperature, are well acknowledged, studies of trends in measures of their variability are much less common. This is despite the fact that trends in variability can have higher effect on climate extremes than trends in mean. Four measures of intraseasonal variability are examined: (a) standard deviation of mean daily temperature, (b) the range between the 90th and 10th quantile of mean daily temperature, (c) mean absolute value of day-to-day temperature change, and (d) one-day lagged temporal autocorrelation. ECA&D daily data from 168 stations and linear regression method are utilized to calculate trends of these characteristics in period from 1961 to 2018. Significant trends (positive and negative) are revealed with substantial differences between seasons, regions and measures. The most considerable decreases in temperature variability were recorded in winter, for temporal autocorrelation in eastern Europe and for variance-based measures in northern Europe. For example, the standard deviation has decreased by more than 10% in the Arctic Ocean. This can indicate a decrease in the frequency of cold extremes in Scandinavia. On the contrary, increasing persistence may suggest a greater likelihood of cold extremes in the East European Plain in winter. Increases in variability prevail only in summer, but not for all measures and not as clearly as decreases in winter. However, the increase in variance-based measures in central and eastern Europe is noticeable. Trends in temporal autocorrelation and day-to-day change appear to be sensitive to data issues, such as inhomogeneities and changes in observational procedure.

How to cite: Krauskopf, T. and Huth, R.: Trends in intraseasonal temperature variability in Europe, 1961 - 2018, EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-539, https://doi.org/10.5194/ems2022-539, 2022.

Onsite presentation
Romana Beranova and Jan Kyselý

Large-scale heavy precipitation events (LHPEs) are associated with negative impacts on society, mainly as they may trigger floods and landslides. Therefore, it is important to better understand underlying physical mechanisms leading to these extremes and how they are reproduced in climate models. The present study evaluates ability of regional climate models (RCMs) to reproduce relationships between LHPEs over the Czech Republic and atmospheric circulation. We use an ensemble of 32 RCM simulations with the 0.11° resolution, taken from the Euro-CORDEX project. The historical simulations (1951-2005) are compared against observations from the E-OBS gridded data set. LHPEs are defined as days with at least 70% of all grid points over a given area with precipitation amounts exceeding the 90th grid-specific percentile of the seasonal distribution of daily amounts. The association with atmospheric circulation is investigated through circulation types derived from sea level pressure using airflow indices (direction, strength and vorticity). The analysis is carried out separately for summer (JJA) and winter (DJF) season.

In observations, the frequency of all LHPEs is higher in winter than summer, which is associated with larger importance of frontal stratiform precipitation in cold part of the year. The links of LHPEs to circulation differ between the seasons and between the western and eastern part of the examined area with varied Atlantic/Mediterranean influences. In winter, observed LHPEs are connected mainly with cyclonic type and westerly supertype. Connection with non-westerly and cyclonic-nonwesterly supertypes is more frequent only in the eastern region. The majority of RCMs simulate differences in circulation-to-precipitation link between the western and the eastern regions well. However, almost all RCMs overestimate the frequency of cyclonic type during LHPEs. In summer, the importance of cyclonic and non-westerly types in producing LHPEs increases compared to winter, and RCMs are able to capture this pattern. The prevailing type during events simultaneously occurring in both regions is cyclonic type with frequency 56%, which is 6 times higher than its mean summer occurrence. Only few models are able to simulate the strong link between events simultaneously occurring in both regions and cyclonic type. In conclusion, the method proposed for LHPEs is able to capture the precipitation events potentially leading to floods and may be suitable to characterize the flood-risk conditions in future climate scenarios.

How to cite: Beranova, R. and Kyselý, J.: Links between large-scale heavy precipitation over the Czech Republic and atmospheric circulation in regional climate model outputs, EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-64, https://doi.org/10.5194/ems2022-64, 2022.

Onsite presentation
Eva Plavcová et al.

Since long-term changes in climate variability are an important aspect of climate change, it is necessary to know whether and how they will change in the future. To this end, it is also important that they are simulated correctly by climate models. Despite its importance to the society and relevance in various impacts, short-term precipitation variability was given very little attention even in the last IPCC report. In our analysis, we aim to answer these questions: (1) What long-term changes in day-to-day precipitation variability have been observed? (2) Do the short-term precipitation variability and its long-term trends differ between various datasets, and how much? (3) How successful are climate models in reproducing precipitation variability? Is it possible to identify biases and errors common to all or majority of models, or to groups of models? (4) What changes of short-term precipitation variability do climate models project in future? Are the observed changes consistent with the projected changes, that is, does the anthropogenic climate change already manifest in the recently observed changes of variability?

We analyze day-to-day variability in precipitation in North Atlantic and European sector. We consider wet-to-wet and dry-to-dry transition probabilities as a measure of short-term precipitation variability, focusing on winter (DJF) and summer (JJA) seasons separately. We compare results obtained from stations data, gridded observed data (E-OBS), various reanalyzes, and outputs of an ensemble of CMIP6 global climate models. We analyze historical model runs starting in 1950 and possible climate change over the 21st century in the simulations under the SSP585 scenarios.

How to cite: Plavcová, E., Huth, R., Beranová, R., and Lhotka, O.: Varying precipitation variability over Europe in observed data and climate models, EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-223, https://doi.org/10.5194/ems2022-223, 2022.

Onsite presentation
André Fonseca and João Santos

Climate variability has a great impact on wine production and quality. To assess weather variability and climatic conditions, this research examines high-resolution extreme indexes of daily temperature and precipitation over 50 protected wine denominations of origin (DOs)/sub-regions in mainland Portugal. This study highlights the variations of present and future climate conditions between DO/sub-regions, for a better understanding of the nature of climate change across diverse regions of the country.

The analysis is based on a high-resolution dataset over mainland Portugal and for a baseline period (1981–2010). Climate change projections are also assessed for two scenarios (RCP4.5 and RCP8.5) and using a 5-member climate model ensemble retrieved from EURO-CORDEX climate projections data over the future periods of 2041–2070 and 2071–2100. A quantile mapping technique was used to bias correct the outputs from general circulation model (GCM) and regional climate model (RCM) combination.

The results highlight the spatial variability of Portuguese wine sub-regions for the present and future conditions. This study also shows that for future periods (regardless of the scenario) the wine sector in Portugal will likely see important climate changes across most regions. Increases in mean temperatures for each region accompanied by increasing severe dryness are projected in future climates mainly in south-eastern Portugal and along the upper Douro Valley (Douro Superior) in north-eastern Portugal. These DO/sub-regions are projected to become much drier than currently so irrigation or the introduction of new varieties are likely adaptation measures to maintain the viability and sustainability of regional viticulture in future decades.

Acknowledgments: Soil recover for a healthy food and quality of life (SoilRec4+Health). Projeto cofinanciado pelo Fundo Europeu de Desenvolvimento Regional (FEDER) através do Programa Operacional Regional do Norte (NORTE-01-0145-FEDER-000083).

How to cite: Fonseca, A. and Santos, J.: Portuguese wine denominations of origin under changing climates, EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-600, https://doi.org/10.5194/ems2022-600, 2022.

Onsite presentation
Matilde García-Valdecasas Ojeda et al.

Drought is one of the most devastating natural hazards that pose significant challenges to human and ecosystems. This kind of phenomenon is expected will become more frequent and severe response to changing climate, especially in regions such as the Iberian Peninsula (IP), which is already being affected by recurring drought events.

One of the most difficult challenges in assessing drought-related impacts is the lack of a universal definition of drought, consequence of the complex nature of this type of extreme phenomenon. That is, a wide range of hydrological variables are involved in the occurrence of drought events leading to different drought types, which may occur at the same time or sequentially. In this framework, multivariate drought indices were developed to address an insufficient characterization of droughts using a single variable.

This study explores drought projections over the IP for different warming levels (1.5, 2, and 3ºC above pre-industrial levels) under the RCP8.5 emission scenario. For this aim, climate data from a multi-ensemble of regional climate simulations within the EURO-CORDEX initiative has been used to compute a set of multivariate standardized drought indices in their nonparametric form. Drought indicators that take into account variables such as the precipitation, soil moisture, and potential evapotranspiration have been computed for different time aggregations. Then, drought characteristics (i.e., duration, frequency, severity, intensity, and spatial extent) from the different indices have been computed in order to compare projected changes in drought conditions for this region.

The findings of this work could provide valuable information to policymakers in developing drought-related adaptation and mitigation policies in response to climate change.

Keywords: Nonparametric multivariate standardized drought indices, drought projections, Iberian Peninsula, EURO-CORDEX regional climate simulations.

Acknowledgments: This research has been carried out in the framework of the projects B-RNM-336-UGR18, funded by FEDER / Junta de Andalucía - Consejería de Economía y Conocimiento, P20_00035, funded by FEDER/Junta de Andalucía-Consejería de Transformación Económica, Industria, Conocimiento y Universidades, by the Spanish Ministry of Science and Innovation (project reference LifeWatch-2019-10-UGR-01), and project CGL2017-89836-R, funded by the Spanish Ministry of Economy and Competitiveness with additional FEDER funds. Fourth author acknowledges the Spanish Ministry of Science, Innovation and Universities for the predoctoral fellowship (PRE2018-083921).

How to cite: García-Valdecasas Ojeda, M., Possega, M., Romero-Jiménez, E., Rosa-Cánovas, J. J., Yeste, P., Castro-Díez, Y., Esteban-Parra, M. J., Di Savatino, S., and Gámiz-Fortis, S. R.: Drought characteristic projections over the Iberian Peninsula: the effect of using nonparametric multivariate standardized drought indices , EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-417, https://doi.org/10.5194/ems2022-417, 2022.

Presentation form not yet defined
Goran Trbic et al.

Intensive urbanization and global warming are impacting the health and well-being of urban population. Nevertheless, urban environments with different design will have different micro and local climate conditions. Therefore, the application of parameters that assess thermal stress in urban areas, such as outdoor thermal comfort (OTC) indices are of paramount importance.

In this study the long-term (1961–2020) datasets of daily OTC indices for the city of Banja Luka (Bosnia and Hercegovina) were used. Detailed temporal analysis using Physiological Equivalent Temperature (PET), Universal Thermal Climate Index (UTCI) and Mean Radiant Temperature (Tmrt) was performed for: a) the entire research period, b) the decadal level, and c) defined extreme heat/cold stress categories. The results show intensive increase in extreme/strong heat days in the last twenty years, and the number of these days is five times higher than in the 70s and 80s. Decreasing tendencies are noticed in extreme/strong cold days towards the last two decades.

We also performed modeling of changes in mean daily air temperature according to IPCC climate scenarios RCP 2.6, RCP 4.5 RCP 8.5 until the end of the XXI century. The results point to significant positive changes under all three climate scenarios.

Furthermore, we analyzed PET, Tmrt and modified Psychologically Equivalent Temperature (mPET) indices during the hot summer days in June 2021 for different urban spaces (downtown, urban park, riverside) in Banja Luka. Results show that the downtown is the most most uncomfortable area in terms of OTC indices. Also, riverside had lower average Taduring summer daytime compared to urban park and downtown likely due to the synergy between river cooling effect (evaporation and sensible heat transfer) and tree shade.

How to cite: Trbic, G., Savic, S., Milosevic, D., Ivanisevic, M., Garic, B., and Markovic, M.: Long-term biometeorological conditions in the mid-sized European city – A case study of Banja Luka (Bosnia and Herzegovina), EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-300, https://doi.org/10.5194/ems2022-300, 2022.

Onsite presentation
Patricia Tarín-Carrasco et al.

Future projections of rainfall changes remain one of the most prominent challenges for climate models due to the high degree of uncertainties associated with them. This is especially true for regions featuring Mediterranean climate, where the uncertainties in projections of future rainfall changes in combination with increasing temperatures, pose a significant problem. One particular uncertainty hotspot is California, where models tend to disagree about the sign of future precipitation changes.

We analyse the simulations from the Coupled Model Intercomparison Projects Phase 5 (CMIP5) and 6 (CMIP6) and study their respective intermodel differences in projections of future rainfall changes over the state of California. We employ 30 models from CMIP5 and 30 models from CMIP6 initiatives and focus on the business-as-usual scenarios (SSP585 - RCP8.5). We consider the precipitation changes over the two time horizons  - mid-century (2040-2059) and the end of century (2080-2099).

Our results point at higher uncertainty in rainfall projections for the wet winter season December-January-February (DJF) than for the rest of the seasons. Comparing CMIP5 and CMIP6, we see that CMIP5 simulations show larger intermodel disagreement for the end-of-century period while CMIP6 show larger intermodel disagreement for the mid-century time horizon. For the new CMIP6 simulations, more than 70% of the models project that northern California is going to experience an increase in precipitation in the future, while the projections for southern California are more uncertain, with models being almost equally spread between those suggesting drier and those suggesting wetter conditions. This is an improvement compared to the CMIP5 simulations where 60% models suggested wetter conditions (and 40 % drier conditions) over northern California.

Finally, we analyse how the uncertainties in future model simulations of geopotential height, outgoing longwave ration (OLR) and ENSO changes relate to the described uncertainties in California’s rainfall in CMIP6 and CMIP5.

How to cite: Tarín-Carrasco, P., Petrova, D., Galí-Reniu, M., and Cvijanovic, I.: Comparison of future rainfall projections over California: CMIP6 vs. CMIP5, EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-619, https://doi.org/10.5194/ems2022-619, 2022.


Onsite presentation
Kelly Stanley et al.

This study investigates whether the recent extremely hot and dry summers (2003, 2015, 2018 and 2019) in Germany will be normal summers under future climate change conditions from climate projections. Abnormally persistent high-pressure systems during these recent German summers maintained clear skies and dry conditions on the ground, resulting in record-breaking heat and drought conditions which had major socio-economic impacts, e.g. reduced crop yields or impacts on human health and increased mortality. Heat and drought indices are calculated from the German Meteorological Service's (DWD) climate model reference-ensembles for periods 2031-2060 and 2071-2100 under greenhouse gas emission scenarios RCP 2.6 (11 ensemble members) and RCP 8.5 (21 ensemble members). These results are compared with the same indicators calculated from three observational datasets (station data and gridded datasets: HYRAS and DWD Climate Monitoring Grids). The order of recent extreme summers in Germany from hottest and driest is: 2018, 2003, 2019, 2015. The climate projections show increasing summer heat conditions for Germany throughout this century, especially for the RCP 8.5 scenario, but no clear future summer precipitation anomaly trend. These recent extreme summers would be considered extremely hot summers for RCP 2.6 in 2031-2060, 2071-2100 and the RCP 8.5 2031-2060 scenarios, but would be considered normal or even below normal hot summers under the RCP 8.5 scenario in 2071-2100. Due to the combination of extreme heat and strong precipitation deficits, the climatic water balance and derived meteorological drought indicator values of these summers often exceeds the considered likely ranges projected for all future summer scenarios and are only matched by the most anomalous hot and dry summers in the model ensembles, even for the high-emission RCP 8.5 scenario in 2071-2100.

How to cite: Stanley, K., Leps, N., and Walter, A.: Recent hot and dry summers in Germany in comparison to climate projections , EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-24, https://doi.org/10.5194/ems2022-24, 2022.

Onsite presentation
Zala Žnidaršič et al.

Frost events that occur after the beginning of the growing season pose great risk for fruit trees, especially after budburst. There have been attempts to model frost risk projections for the 21th century in Slovenia, however the results have shown no change in frost risk due to climate change, despite observations for years 2019 and 2020 showing otherwise. Therefore, an alternative climatological analysis of frost event probability is of great importance, especially in combination with phenological data from Slovenian fruit tree orchards and vineyards. An assessment of the risk of tardive frosts for a series of Slovenian fruit tree orchards and vineyards throughout the 21st century was made from temperature data of climate model projections, specifically from regionally downscaled projections data. The analysis bases on two pheonological models for calculation of budburst - a classic GDD model and a two-phased BRIN model. Afterwards the approach consists of a comparison of the statistical probability of two events, the last frost event probability (day of the year) and a budburst event probability, as simulated by the different phenological models. We discovered that the probability of a tardive frost, i.e. frost events occurring after grapevine budburst, in Slovenian grapevine varieties varies throughout the 20th and 21th century, for example, the probability of a tardive frost increases towards the end of the 21th century for Chardonnay (Eastern Slovenia with continental climate) and Riesling (Northeastern Slovenia with continental climate) varieties but decreases for Merlot (Southwestern Slovenia with Mediterranean climate). An analysis of tardive frost projections for cherry, apple, olive and other fruit tree varieties grown in Slovenia is under way. Some preliminary results of the project will be represented.

How to cite: Žnidaršič, Z., Sušnik, A., Gregorič, G., and Pogačar, T.: Risk assessment of frost occurrence in a changing climate, EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-51, https://doi.org/10.5194/ems2022-51, 2022.

Onsite presentation
Ksenija Cindric Kalin et al.

Observed changes in the seasonal and annual air temperature and precipitation amounts, and the corresponding indices of extremes in Croatia are analysed. Daily data from the 1961-2020 period are employed at 37 climatological stations (mean and extreme air temperatures) and 164 rain gauge stations (precipitation amounts) operated by the Croatian Meteorological and Hydrological Service. The trends are estimated using linear regression and the statistical significance of trends is tested by the Mann-Kendall test. The results revealed a consistent warming trend in the whole country. The highest increase in the mean annual air temperature (up to 0.5 °C/10years) is detected in central continental Croatia mostly due to summer and winter positive trends. Observed warming is also evident in positive trends of warm temperature indices (warm days, warm nights and warm spells) and negative trends in cold temperature indices (cold days, cold nights and cold spells). A significant decrease in the summer precipitation amounts is found along the Adriatic coast and in the highlands (5-15 %/10years), in the latter region also in spring. The most prominent feature of precipitation changes is a consistent positive trend in autumn rainfall amounts across the whole country, with the largest increase detected in central parts. The observed trends are associated with an annual increase in dry spell (DS) lengths in the central parts and along the Adriatic coast, and with a significant decrease in DSs in easternmost Croatia. Moreover, in the central, mountainous and northern Adriatic region, a significant increase in precipitation fraction due to very wet days (R95T) and in daily intensity index (SDII) is found. The summer drying is associated with a decrease in the number of moderately wet days along the Adriatic and the highlands but also with an increased number of dry days and DS lengths, and a decrease in the SDII and highest 1-day (Rx1d) and 5- day (Rx5d) precipitation amounts. On the other hand, autumn wetter conditions are mainly associated with an increase in R95T and with shorter DS lengths but also with an increase in SDII and Rx1d and Rx5d. The spring trend pattern is characterised by longer DS lengths in the northern Adriatic and longer wet spells in eastern Croatia.

How to cite: Cindric Kalin, K., Patalen, L., Marinovic, I., and Pasaric, Z.: Trends in temperature and precipitation indices in Croatia, 1961-2020, EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-212, https://doi.org/10.5194/ems2022-212, 2022.

Onsite presentation
Lucija Blašković et al.

National hail climatology of Croatia


Lucija Blašković, Damjan Jelić, Barbara Malečić, Branimir Omazić, Maja Telišman Prtenjak (1)


(1) University of Zagreb, Faculty of Science, Department of Geophysics, Zagreb, Croatia


The step every country should take towards a better understanding of hail is national climatology. It is crucial to explore the temporal and spatial characteristics of this small-scale phenomenon in hope that it will contribute to better defense and preparation for such extreme weather. Even though Croatia is a small country, it has a relatively inhomogeneous climate due to its proximity to the Adriatic sea, a mountain range that in a way separates Croatia, etc. This research will analyze the diversity of the Croatian climate and its impact on the occurrence of hail. The results are based on hail data from 186 stations. This digitized data vary from station to station, so different time periods were observed. The results showed significant interannual and spatial variability, due to which it was necessary to divide into subdomains based on the season in which hail predominates - on the entire coast, the highest hail activity is present in the colder part of the year while the continental part of Croatia shows increased summer hail activity. There is also a transitional area that records most hail in spring and fall. The trend analysis was made for two time periods: 1.) from 1964. to 2019., 2) from 1984. to 2019. and both periods have shown signs of a negative trend in a number of hail events. Daily patterns show a shift in the daily maximum from morning to afternoon hours as we approach the continent from the coast, and the highest frequencies are recorded in Šibenik and Zavižan of as much as 3.8 and 4.8 days with hail per year. Finally, instability indices were studied, which could explain the atmospheric conditions in which hail occurs.



How to cite: Blašković, L., Jelić, D., Malečić, B., Omazić, B., and Telišman Prtenjak, M.: National hail climatology of Croatia, EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-349, https://doi.org/10.5194/ems2022-349, 2022.

Onsite presentation
Marta Martinkova and Jan Kysely

The Clausius-Clapeyron (CC) equation describes the change in saturated water vapor pressure with temperature (7% per degree C within the normal temperature range). The possible effects of deviations from CC scaling, especially the so-called "super CC" scaling (increase above 7%), on the occurrence and intensity of precipitation extremes are connected to many different factors. The most important of them are the availability of moisture and more intense convection. The resulting CC scaling is influenced also by the type of precipitation and the super CC scaling may be the result/artifact coming from combination of precipitation of different types. Here, we present a study of the variability in the rates of increase in precipitation intensity with dew point temperature using the Multi Source Weighted Ensemble Precipitation (MSWEP) 3h data with the resolution of 0.1 deg. for 2006–2016 over Central Europe. The scaling is assessed for daily dew point temperature (taken from ERA5-Land dataset) and the 95th percentile of 3h precipitation using the quantile regression method. As the resulting scaling may be influenced by the type of precipitation (convective versus non-convective), we apply the method of threshold precipitation rate (THR) to categorize the precipitation in two subsets. The average scaling slopes are 8.7 %/°C for all precipitation events, 6.8 %/°C for precipitation events above THR and 3.0 %/°C for events bellow THR. The mixing of events bellow/above THR amplifies the scaling. To evaluate the scaling of convective precipitation with dew point temperature independently, we use the data of lightning occurrence from the EUCLID dataset (3h, resolution 0.1 deg.), as the lightning occurrence is assumed to be a robust proxy for convective precipitation. The average scaling slopes are 13.4 %/°C for precipitation events with lightning and 6.3 %/°C for precipitation events without lightning. The scaling of the events with lightning corresponds with the double CC scaling for convective events reported by previous studies.

How to cite: Martinkova, M. and Kysely, J.: Precipitation-Temperature Scaling of 3h Convective Precipitation over Europe , EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-88, https://doi.org/10.5194/ems2022-88, 2022.

Onsite presentation
Sarah Ivušić et al.

The study aims to estimate the future climate change of extreme precipitation over the topographically complex coastal-mountainous region of the eastern Adriatic and Dinaric Alps, which is particularly vulnerable to climate change. A number of studies classify this region as an area with a "zero-change" line between the wetter north and drier south, which shifts northward towards the end of the century. However, the research on future extreme precipitation changes over this region is still limited. We use an unprecedented ensemble of ~140 regional climate model (RCM) simulations of future climate from the EURO-CORDEX ensemble at 0.11° resolution, to cover as many future conditions and sources of uncertainty as possible. The ensemble is comprised of 15 RCMs driven by 11 CMIP5 global climate models. The climate change signal is estimated for three different greenhouse gases concentration scenarios (RCP2.6, RCP4.5 and RCP8.5) and several future periods (2041-2070, 2071-2100) with respect to the historical period (1971-2000). We focused on heavy precipitation measures: the 99th percentile of all-day precipitation, number of heavy and very heavy precipitation days, maximum one-day and five-day aggregated precipitation sum. Additionally, we have applied the extreme value analysis, specifically the generalized extreme value theory, to assess extreme precipitation return levels associated with return periods between 10 and 100 years. The results are highly dependent on period, scenario, season and location. Overall, results show an intensification of both heavy and extreme precipitation events, especially during cold seasons over the north-eastern areas for the far future period. For this projected change, models show high agreement as opposed to that in summer, when most of the examined indices display the aforementioned south-north gradient. A more detailed analysis is planned to quantify the climate change signal for several subdomains of interest.

How to cite: Ivušić, S., Güttler, I., and Horvath, K.: Projected extreme precipitation changes over the coastal-mountainous region of the eastern Adriatic and Dinaric Alps, EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-446, https://doi.org/10.5194/ems2022-446, 2022.

Onsite presentation
Péter Szabó et al.

While anthropogenic global warming is well-accepted and thoroughly described in the climate community, the vast majority of the general public is still not always fully-certain whether to associate specific local weather or climate event with this global phenomenon. It is of utmost importance to raise public awareness by providing easily understood but scientifically sound information, and to explain how the anthropogenic activity contributes to specific weather or climate events. Although it is getting to be widely assessed now in Western Europe, only a few case studies for past high-impact local weather events were assessed in Hungary, and the systematic analyses of long-term trends of such indicators with their linkage to anthropogenic activity are missing in this region. Our local attribution project (started in September 2021) aims to fill this gap in Hungary: the results of the analyses of seasonally relevant indices are published in each season at around a time of an event occurrence or absence. Dissemination is executed via an already established, online Hungarian platform (https://masfelfok.hu/, meaning “1.5 degrees”), since it is reaching the general public with its readily understood climate change information through its broad media coverage and a growing social media network.

The analysis uses several freely-available data sets: (1) an ensemble selection of CMIP6 global climate model simulations with both natural-only forcings and historical runs, (2) a fit-for-purpose ensemble of regional climate model simulations from the Euro-CORDEX on 0.11° resolution, including the RCP2.6, RCP4.5 and RCP8.5 future scenarios, (3) a fine-resolution, homogenized observation-based gridded data for Hungary, compiled by the Hungarian Meteorological Service, (4) the newest E-OBS database for the area outside of Hungary. Within the project, we selected seasonally relevant extreme, less extreme but well-known events, and these pre-selected indices are attributed to anthropogenic activity through their detected intensity, duration, or frequency changes. For instance, third-level heatwave days, hot days, tropical nights, annual temperature maxima and return values, meteorological droughts, and heavy precipitation days are evaluated for summer, as they have the largest public interest and impacts in our target region.

How to cite: Szabó, P., Bartholy, J., and Pongrácz, R.: Attribution of pre-selected climate indicators in summer over the Carpathian basin, EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-450, https://doi.org/10.5194/ems2022-450, 2022.

Onsite presentation
Svenja Szemkus et al.

In the BMBF project climXtreme CoDEx, we analyze and explore methods for data compression to detect and attribute changes in the frequency and severity of extreme weather events with respect to climate change. Particularly for local processes on the atmospheric mesoscale, climate change signals are often masked by additional variability, resulting in poor signal-to-noise ratios. Appropriate reduction of degrees of freedom can improve the signal-to-noise ratio, increasing the potential to detect less strong signals in this setting. Our focus is on unsupervised learning approaches such as principal component analysis that are specifically designed for extremes.

We focus on extreme precipitation in Germany and analyze how different data compression techniques can be used in a detection and attribution (D&A) study. Among others, we use the approach of Cooley and Thibaud (2019), who decompose the pairwise dependence matrix as an analogue to the covariance matrix for extreme dependencies. We use principal component analysis and find an index that accounts for precipitation extremes due to common extremes in the first principal components. We use this index for a D&A study.

We further suggest that changes in extreme precipitation may depend on the spatial scale, e.g., when thinking about stratiform and convective precipitation events caused by different atmospheric conditions. Therefore, separate investigation of different event types might be necessary for robust and reliable D&A. This motivates our approach to filter different event types using the dual-tree wavelet transform, as previously applied by Buschow and Friederichs (2021). We propose a scale-dependent D&A study based on the filtered precipitation data.

How to cite: Szemkus, S., Friederichs, P., Buschow, S., and Pscheidt, I.: Climate change detection and attribution in extreme precipitation using compact representations , EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-460, https://doi.org/10.5194/ems2022-460, 2022.

Online presentation
Stephen Outten and Stefan Sobolowski

Every year, extreme wind events cause a great deal of financial losses across Europe. As such, the accurate projection of the changes in extreme winds will be invaluable for many industries, including insurance, construction, energy, and afforestation. Adaptation planning requires the estimation of how the magnitude and frequency of extreme events, such as extreme winds, will vary in the future.

Extreme winds are examined in a selection of Euro-CORDEX downscalings. Since these include multiple downscalings of the same global climate models by different regional climate models, it is possible to intercompare how different models represent extreme winds. The peaks-over-threshold approach is used identify the extreme events based on the fitted Generalized Pareto Distributions (GPD), and maps of 30-year return winds for all locations over Europe are presented. Future changes in the frequency of extreme winds are assessed for Northern, Central, and Sothern Europe for three future periods, being the near, mid, and end of this century. The results show that the frequency of extreme wind events will increase from one period to the next over the 21st century in all three regions of Europe.

The results of the recently published study will be presented in detail along with a brief overview of the statistical methods employed. Our ongoing work to create a comprehensive dataset of assessed extreme winds and precipitation will also be addressed. Time permitting, we will present evidence of the benefits of downscaling, an assessment of how the base period length affects the uncertainties, and a brief comparison of the different methods commonly employed to assess uncertainties.

How to cite: Outten, S. and Sobolowski, S.: Assessment of extreme winds over Europe in Regional Climate Models, EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-365, https://doi.org/10.5194/ems2022-365, 2022.

Onsite presentation
Josh Dorrington et al.

The regional impacts of anthropogenic climate change over Europe are still very unclear, in large part due to uncertainties in the dynamical response of the large-scale circulation. Particularly in Boreal winter, the structure of the Euro-Atlantic circulation is strongly nonlinear, and is dominated by Rossby wave breaking, persistent atmospheric blocks and latitudinal meandering of the eddy driven jet stream. In the face of this complexity, adopting a regime approach – where the large scale circulation is decomposed into a small number of representative flow state – can shed much needed light on the internal variability and forced response of Euro-Atlantic dynamics.


One major open question is how exactly nonlinear regimes respond to a warming climate. While it has been hypothesised that climate forcing should primarily result in a shift in the probability of historically observed regimes, rather than inducing new flow configurations, this has never been tested in complex earth system models. In this work we fill this gap, by analysing regime statistics in 20 CMIP6 models under the SSP5-8.5 scenario. We decompose the forced response into two components: one representing the degree to which the prevailing regime patterns alter, and another representing their changing probability of occurrence and persistence. We show that while robust changes in regime patterns can be detected, these are minor in comparison to changes in the regime life-cycles. The overall picture tends towards a less regime dominated future: we find that anticyclonic regimes associated with persistent blocking are projected to become less persistent in the future and that central jet positions, associated with zonal flows, become more common. However considerable intermodel variability, even in the sign of the prevailing regime trends, shows the need for further investigation, and an improved representation of Euro-Atlantic regimes in state-of-the-art models.

How to cite: Dorrington, J., Strommen, K., Fabiano, F., and Palmer, T.: How do Euro-Atlantic regimes respond to climate change?, EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-266, https://doi.org/10.5194/ems2022-266, 2022.

Onsite presentation
Future summer changes in zonal surface circulation types and mid-tropospheric weather regimes over Europe
Meriem Krouma et al.
Onsite presentation
Anita Verpe Dyrrdal et al.

The Nordic-Baltic region has experienced numerous flooding episodes resulting from heavy rainfall. Such events are costly and may potentially threaten the safety of the population. As part of the cooperation in the Nordic Framework for Climate Services (NFCS), we here present results from two separate analyses of heavy precipitation in a large region covering Fennoscandia and the Baltics (Dyrrdal et al., 2021; Olsson et al., 2022).

In Dyrrdal et al. (2021), long-term (1901–2020) changes of annual maximum daily precipitation from 138 stations and short-term (1969–2020) changes from 724 stations were analyzed, along with their date of occurrence. Results show a majority of positive trends in daily annual maxima and the 5-year return level, with hotspots in southeast Norway, southern Sweden and southwest of Finland. Almost all stations with a statistically significant trend showed an increase in the intensity of annual maximum 1-day precipitation during the last 50 years. In the region as a whole, annual maximum precipitation events occur somewhat later in the year now compared to the beginning the last century

For urban flooding, the occurrence of short-duration rainfall is of special importance. In Olsson et al. (2022), sub-daily rainfall observations from 543 meteorological stations in the Nordic-Baltic region were collected, quality controlled and consistently analyzed in terms of records, return levels, geographical and climatic dependencies, time of occurrence of maxima and trends. The results reflect the highly heterogeneous rainfall climate in the region, with longitudinal and latitudinal gradients as well as local variability, and overall agree with previous national investigations. Trend analyses for Norway and Denmark indicated predominantly positive trends in the period 1980-2018. 



Dyrrdal, A.V., Olsson, J., Médus, E., Arnbjerg-Nielsen, K., Post, P., Aņiskeviča, S., Thorndahl, S., Førland, E., Wern, L, Mačiulytė, V. and Mäkelä, A., 2021: Observed changes in heavy daily precipitation over the Nordic-Baltic region, Journal of Hydrology: Regional Studies, 38, 100965, ISSN 2214-5818, https://doi.org/10.1016/j.ejrh.2021.100965.

Olsson, J., Dyrrdal, A.V., Médus, E., Södling, J., Aniskeviča, S., Arnbjerg-Nielsen, K., Førland, E., Mačiulytė, V., Mäkelä, A., Post, P., Thorndahl, S.L., and L. Wern (2022) Sub-daily rainfall extremes in the Nordic-Baltic region, Hydrol. Res., in press.


How to cite: Dyrrdal, A. V., Olsson, J., Médus, E., Arnbjerg-Nielsen, K., Post, P., Aņiskeviča, S., Thorndahl, S., Førland, E., Wern, L., Mačiulytė, V., Mäkelä, A., and Södling, J.: Observed changes in heavy precipitation over the Nordic-Baltic region, EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-368, https://doi.org/10.5194/ems2022-368, 2022.

Onsite presentation
Rovina Pinto and Stephanie Fiedler

The Atacama Desert should be a region prone to wind erosion and dust events given the barren landscape and lack of moisture but it rarely sees strong dust events. Despite this perception, there is no literature quantifying the frequency of dust events in the Atacama and analysing the underlying meteorological mechanisms responsible for dust storms in the region. To this end, we analysed surface observations spanning 37 years (1984-2021) to compile wind climatology and systematically assess the dust events in the Atacama Desert. We used the Met Office Integrated Data Archive System (MIDAS) Land and Marine Surface Stations Data. Based on the present weather codes for nine Chilean stations present in the data, frequency of dust events was computed. The Atacama Desert has a very low annual mean dust event frequency of 0.6% despite its hyper-aridity. The station at Chañaral records the highest number of dust events, accounting for 68% of all recorded dust observations. There is no perceptible seasonal pattern but there is a strong diurnal cycle in the dust activity in the Atacama as most dust events occur between 1500 and 1800 local time. Threshold wind speeds, defined as the minimum wind speed required for 5, 25 and 50% of dust event frequency, are estimated for all stations except at Vallenar. The thresholds allow us to determine the lowest winds capable of emitting dust from the surface and also the highest wind speeds required, and the differences between the thresholds is indicative of the changes in soil conditions such as soil moisture or land use. For 5% of dust event frequency, the lowest threshold wind speed of 6-8 ms-1is estimated at Chañaral and for 50% of dust event frequency, 10-12 ms-1at Desierto de Atacama. Higher threshold wind speeds are observed for stations north of 26°S than the stations south of 26°S. It implies that dust particles can be mobilized by weaker near-surface wind speeds in the South than in the North of the Atacama. Seasonal variations in the threshold wind speed are not strongly pronounced in stations where sufficiently many observations are available. Ongoing work includes incorporating additional datasets for filling in identified data gaps in MIDAS and for allowing a spatially broader analysis of dust-emitting winds across the Atacama Desert.

How to cite: Pinto, R. and Fiedler, S.: Dust Event Assessment for the Atacama Desert using Surface Synoptic Observations: 1984 - 2021, EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-90, https://doi.org/10.5194/ems2022-90, 2022.

Onsite presentation
Gudrun Nina Petersen

With increasing air temperature in a warming climate the soil temperature also increases. Such changes impact flora, fauna and natural hazards. Thus, it is important to monitor soil temperature. In Iceland, the highland station Hveravellir has the longest reliable time series of soil temperature. The station is located in a barren landscape at 641 m a.s.l., but the site itself is grass covered. Soil temperature has been measured there since 1965 at 10, 20, 50 and 100 cm depth but the data is not fully digitized. Here, the time series used span 42 years, from 1977 to 2019. Annual and seasonal trends are calculated as well as the annual number of thawing degree days (TDD) and freezing degree days (FDD). During this period, there has been a considerable warming in Iceland, not only because of global warming but also because of its start coinciding with the end of a local cold period. In the soil, warming is detected in all seasons except May to June, which is the melting season, the end of which varies greatly from year to year. The most warming is found during the autumn and winter months. These results contrast with those from continental cool sites where the largest warming has been seen in the spring months. The autumn cooling starts 2–3 weeks later now than at the start of the time series, resulting in a longer summer in the soil. The number of TDD has increased and the number of FDD has decreased, both in air and in soil. At 100 cm depth TDD has increased by 80°C day dec −1 and FDD decreased by 38°C day dec −1. There is a correlation between the increase in air temperature and soil temperature, for every 1°C increase in 2 m temperature the soil temperature at 100 cm has increased by 0.6°C. The implication of this warming is firstly seen in terms of flora, fauna and agriculture, by the lengthening of the growing season, and secondly in terms of natural hazard, by increased risk of landslides due to thawing of permafrost in montainous regions.

How to cite: Petersen, G. N.: The soil warming in the Icelandic highlands, EMS Annual Meeting 2022, Bonn, Germany, 5–9 Sep 2022, EMS2022-588, https://doi.org/10.5194/ems2022-588, 2022.

Supporters & sponsors