Description of work

Wp2 will produce new estimates of UQ and, where appropriate, PDFs for future climate on projection time- scales out to 40+ years, building on previous assessments, for instance in the UK, Australia, and Switzerland using the latest climate models and more recent consideration of suitable constraints from other Horizon 2020 projects, such as CRESCENDO, APPLICATE and PRIMAVERA. Additionally, we will produce a range of plausible realisations of future climate and extreme weather (building on progress in, e.g., The Netherlands).

This WP will draw climate data in (for instance from CMIP6), and assess/evaluate historic and present day performance using observations and regional reanalyses (from C3S). A major part of the work will then be to develop and apply advances in applying emergent constraints to these model simulations, including those constraints that emerge from attribution approaches. This first aspect can be summarised as making use of observations, ensembles of 20th Century simulations, and re-analyses to constrain the models and generate a UQ/PDF of future response. The second aspect will then be to place existing and a small number of new simulations in the context of the PDFs, including providing realisations of future climate for use directly or with further downscaling, for instance in WP3. We will look at results on a seasonal and sub-seasonal time resolution.

Objectives

This work package will improve the methods used to characterise uncertainty in future projections of climate change and variability out to around 40+ years, and to develop and test methods to provide new realisations of future climate within the envelope of uncertainty. As such, we have the following objectives:

  • Test existing and develop new emergent/observational constraints in order to produce uncertainty distributions/probability density functions (PDFs) of future climate change.
  • Apply these constraints to produce uncertainty quantifications/PDFs of future climate change, and compare with other traditional methods (unweighted model results, Reliability Ensemble Averaging)
  • Determine contributions of natural variability, model uncertainty and scenario uncertainty to the provided uncertainty quantifications.
  • Produce a limit set of climate scenarios (“future climate”) and extreme weather situations (“future weather”) that sample the PDFs in a number of user relevant statistics, and provide these as boundaries for further downscaling with non-hydrostatic model (for WP3).