S-CMIP: Swedish climate research and contributions to the sixth International Coupled Model Intercomparison Project (CMIP6) - phase 3 (2017)
Since 1995, the Coupled Model Intercomparison Project (CMIP) has internationally coordinated climate model experiments on behalf of World Climate Research Programme (WCRP). By documenting, analyzing and comparing, insights are gained into climate processes, variability and climate change. In this proposal, the Swedish community of Earth System Modeling, with members from Stockholm, Lund, Gothenburg and Uppsala Universities together with SMHI, is asking for computational resources for climate science and earth system projects related to the sixth phase of CMIP (CMIP6) and to ongoing research projects leading to improved capabilities in CMIP6. CMIP6 has started and is planned to end not before 2020. CMIP generates a wealth of scientific value and also provides the basis for assessment of regional and local climate by downscaling of the global scale information to finer scale. The recent COordinated Regional climate Downscaling Experiment (CORDEX) is the primary example for regional downscaling of CMIP information. The combined results of CMIP lead to improved decision support, and allow for informed measures targeting climate change mitigation and adaptation. The primary tool of S-CMIP is the EC-Earth model. It is a state-of-the-art Earth System Model (ESM), developed and used by a consortium of European Met Services, research institutes and universities from 12 European countries. S-CMIP aims at a further consolidation of the development of different EC-Earth model configurations including component model for ocean bio-geo-chemstry and dynamic vegetation. Furthermore, studies are planned concerning climate processes, climate prediction, climate sensitivity, climate projection and consequences of a warmer global climate, all of those preparing for, or being part of, the international CMIP6 effort. In particular, S-CMIP proposes computations in 4 areas in 2017: 1) Process studies and model improvement 2) CMIP6 tuning and spin-up simulations 3) CMIP6 DECK, core simulations characterizing the Earth System Model 4) CMIP6 Model Intercomparison Projects (MIPs) For climate simulations leading to CMIP6, and for the actual CMIP6 simulations as described above, we currently estimate an overall remaining amount of 200 million core-hours (estimates based on Triolith units) over a period of the remaining 3 years (2017-2019). As multi-year allocations are not possible in SNIC, we propose an allocation of 55 million core-h for 2017 and propose to provide updated allocations for the remaining period every year to SNAC. Climate simulations of the proposed type generate substantial amounts of data, which needs to to be stored, processed and published for the international science community. Those tasks require a data infrastructure, which resides on e.g. large accessible disks, which exist at NSC. Therefore, appropriate links between the compute nodes and the data infrastructure are essential.