The new generation of high resolution climate models, known as convection permitting regional climate models (CPRCM), reproduce precipitation extremes much more realistically compared to the regional and global climate models (GCMs) used until now. They also show larger increase in extreme precipitation for future climate compared to lower resolution models. Additionally, due to their high resolution, they have significant impact on climate projections of urban temperature, climate over mountainous regions and wind over complex coastlines.
Uncertainty quantification as part of climate information is essential to establish best possible climate adaptation strategies for decision makers. Uncertainty estimates are obtained from large ensembles of climate simulations, which are typically not feasible for convection permitting models due to their high computational costs. We make use of the first-ever international ensembles of convection permitting simulations to systematically investigate the dependence of uncertainty range on ensemble size, and its transferability to different regions. We will incorporate a clustering approach and high user engagement to develop an optimized approach for reducing ensembles with minimized loss of uncertainty information.
The aim of EDUCAS is to find optimal approaches for minimising CPRCM ensemble size while ensuring that the climate change information and uncertainty are adequately represented for different users.
Since only a few GCMs can be downscaled by CPRCMs, the choice of GCMs has to be done systematically and efficiently. The specific objectives of EDUCAS are:
- Find the smallest set of GCMs that adequately represent internal climate variability for specific user needs (e.g. range of possible changes in urban thermal comfort and extreme precipitation for urban planners).
- Investigate how downscaling with CPRCMs affects the climate variability for specific users. E.g. there are early indications that CPRCMs reduce the uncertainty in projections of precipitation due to better treatment of convective processes.
- Quantify how the decrease in GCM-CPRCM ensemble size affects the climate change uncertainty estimation. What is lost if certain GCM-CPRCM combinations are removed randomly from the existing larger ensembles? What is the optimal way to reduce the ensemble size for different users?
- Find the best practices for choosing which GCMs to downscale for future CPRCM projects.
EDUCAS is funded by FORMAS.
EDUCAS will run from January 2020 to December 2022.
Contact person at SMHI