There is a wide range of climate projections available for Europe. For long-term climate projection information, several global (CMIP3/5/6) and regional (EURO-CORDEX, or the ENSEMBLES project) climate model datasets are available. In addition, a number of EU countries have their own national climate projections (such as UKCP18 for the UK, KNMI-14 for the Netherlands, or CH2011 for Switzerland). This can make it challenging for users to navigate and choose from this variety of projections. Particularly, different climate models will often give different projections of future climate. The range of possible outcomes given by a group of models gives a measure of climate model uncertainty. In order to better inform decision makers, it is important to better understand and measure this uncertainty in climate projections. Assessments have already been carried out to better understand this, for instance in the UK, Australia and Switzerland, using the latest climate models, and more recent research from other Horizon 2020 projects, such as CRESCENDO, APPLICATE and PRIMAVERA.


Work Package 2 (WP2) aims to extract more useful information from the diverse and complex landscape of climate projections in order to make it easier for users to include them in their decision-making process. WP2 does this in two main ways.

Firstly, WP2 looks at developing methodologies to better understand and measure the uncertainty in climate projections (also called ‘characterising uncertainty’). The aim of this evaluation is to highlight which parts of the projections’ range are most and least plausible. WP2 does so through:

  • Testing existing and developing new constraints on climate model outputs for Europe on timescales out to 40+ years, using observations and models. This is then used to produce uncertainty quantifications (UQs, that measure uncertainty and its underlying causes) and probability density functions (PDFs) which describe the most likely pathways for future climate change.
  • Comparing this approach with other traditional methods such as unweighted model results and Reliability Ensemble Averaging, which analyses the performance of the model in reproducing present-day climate and the convergence of the simulated changes across models.
  • Determining the contributions of natural variability, model uncertainty and scenario uncertainty in the uncertainty quantifications.

Secondly, an alternative and complementary way to extract more useful information from future climate projections is through ‘storylines’.  Alongside the quantified uncertainty ranges, WP2 also looks at a range of approaches to identify plausible scenarios of future weather and climate relevant to a number of users. These may be placed in the context of the PDFs described above or might be used on their own as part of a robust decision-making approach.

Read more about activities