Major uncertainties remain in our understanding of the processes that govern the water cycle in a changing climate and their representation in weather and climate models. Of particular concern are heavy precipitation events of convective origin (i.e., thunderstorms and rain showers).

The development of weather and climate models has made rapid progress in recent years. With the progress of high-performance computing (HPC), the computational resolution of such models will continue to be refined in the next decades. This development offers exciting prospects. From a climate science perspective, a further increase in resolution will make it possible to base such models on a set of equations that is much closer to first principles. In particular, at horizontal resolutions of a few kilometers, the models start to explicitly represent the dynamics of deep convective and thunderstorm clouds without the help of semi-empirical parameterizations. This development corresponds to an important quantum jump in climate modeling. It allows reducing some of the key uncertainties in the current generation of climate models, yields an improved representation of the water cycle including the drivers of extreme events (heavy precipitation events, floods, droughts, etc.), and enables more sophisticated climate-change scenarios with better guidance for impact assessment and climate change adaptation measures.

From a computer science perspective, this strategy poses major challenges. First, emerging hardware architectures increasingly involve the use of heterogeneous many-core architectures consisting of both CPUs and accelerators (e.g., GPUs). The efficient exploitation of such architectures requires a paradigm shift and has only just started. Second, with increasing computational resolution, the models’ output becomes unbearably voluminous and long-term storage prohibitively expensive. Ultimately, there is no way around conducting the analysis online rather than storing the model output, and performing model reruns (i.e., repeat simulations for refined analysis). These developments pose new challenging computer science questions, which need to be addressed before an efficient exploitation of new hardware systems becomes feasible.

The project crCLIM is organised in 4 subprojects with tight interactions:

Subproject A (Lead: O. Fuhrer): Assesses how to efficiently exploit heterogeneous many-core computing architectures for weather and climate models. This involves addressing key questions such as the relative costs and benefits of higher-order numerical accuracy versus higher spatial resolution.

Subproject B (Lead: T. Hoefler): Explores the virtualization of climate simulations from a computer science -perspective. It addresses the critical question of computational versus mass-storage loads, and develops an online analysis platform for high-resolution models.

Subproject C (Lead: C. Schär): Performs and analyzes convection-resolving climate simulations over Europe. The main objectives are to demonstrate the feasibility of European-scale climate simulations at convection-resolving resolution, and to exploit this methodology to improve the understanding of the European water cycle.

Subproject D (Lead: H. Wernli): Exploits the new level of virtualization using Eulerian and Lagrangian perspectives for the online analysis of synoptic features and water transport.