Studies on the impacts of climate change require robust climate change data at the local level to simulate the potential sectoral impacts of climate change. However, impact models can often only handle a limited number of climate simulations. A practical sub-selection of climate ensemble members is therefore frequently helpful and has been developed at MeteoSwiss.
Climate model projections show how the climate might evolve under different emission scenarios. As each individual climate model is only a simplified representation of reality, different models produce different results even when identical input conditions are specified (model uncertainty). This uncertainty can be narrowed down and quantified by using several climate models – a so-called climate model ensemble.
To investigate the sectoral impacts of climate change, the climate model ensemble can be integrated into impact models. However, such impact models are often only able to handle a limited number of climate simulations. Selecting ensemble members that are representative of the specific case can help to address this issue.
To meet these user needs, MeteoSwiss and ETH Zurich have developed a method for subsampling members from a climate model ensemble of the CH2018 climate scenarios (Sikorska-Senoner et al. 2024). The methodology is based on initial analyses conducted as part of the CH2018 climate scenarios (Strassmann and Fischer, Chapter 10.4 in the CH2018 Technical Report). The ensemble sub-selection developed on the basis of CH2018 has now been updated for the new Swiss climate scenarios, Klima CH2025 (MeteoSwiss and ETH Zurich, 2025a).
The method selects ensemble members based on the strength of the climate change signal (CCS), which has been calculated for a range of climate indicators. For each application and each GWL
(Global Warming Level) three representative ensemble members are selected, representing the median (50th percentile), the strongest (95th percentile) and the weakest (5th percentile) CCS. This sub-selection allows the spread of the complete climate model ensemble to be represented to a certain extent. In principle, the use of the complete ensemble is recommended. The sub-selection should only be used if this is not possible.
Users can choose between seven predefined application cases covering the whole of Switzerland, as well as the five major regions of Switzerland and various geographical areas. The simulations or ensemble members selected via the sub-selection can be used for both the DAILY-GRIDDED and DAILY-LOCAL datasets from Climate CH2025. The pre-calculated tables and graphs are available in the CH2025 Web Atlas, i.e. CH2025: Ensemble Sub-selection. A sub-selection for an expert case is also available on request.
The detailed methodology of the original ensemble sub-selection based on the earlier CH2018 climate scenarios is described in Sikorska-Senoner et al. (2024). A summary of the ensemble sub-selection, updated for the CH2025 climate scenarios, can be found in «Guide for users».
Users of the Swiss climate scenarios currently have seven predefined use cases at their disposal. There are initially four applications covering the key messages of Klima CH2025: drought, heatwave, heavy rainfall and winters with little snow. These are supplemented by three further applications that illustrate climate change using the most important basic climate variables: temperature, precipitation, and a combination of temperature and precipitation. If there is any uncertainty regarding the choice of an application, it is recommended to select the overarching ‘temperature’ case, which is based on five temperature indices calculated on an annual and seasonal basis.
The expert case is currently only available on request by email to klimaszenarien@meteoswiss.ch. Personal communication between users and MeteoSwiss is necessary to define the specific use case optimally and to enable the best possible sub-selection of ensemble members. Feedback from users regarding current or potential future use cases for the sub-selection, sent to the email address, is very welcome and will be systematically collected to expand the available range of use cases.
User feedback can be sent directly to the following email address:
klimaszenarien@meteoswiss.ch
