ALBERT

Description: A global kilometer-scale (K-scale) “convection-resolving” model (GCRM) simulation on the climate time scale is one of strategies for the more accuracy of cloud-related processes in the climate system. Against this background and the recent advent of a flagship supercomputer “Fugaku” in Japan, a reliable global K-scale multi-year simulation using the Nonhydrostatic ICosahedral Atmospheric Model (NICAM) is now in sight. While this activity expects the good reproducibility of both climatological statistics and weather disturbances, we also know that this is not straightforward in that there exist resolution-independent model errors at least for O(1–10)-km grid spacing. We have tackled this problem, which partly comes from the poor constraint of a moisture-convection relation contributed by physics-dynamics coupling even in K-scale, via the update and/or tuning of cloud microphysics, unresolved turbulence, and vertical resolutions. In an ongoing AMIP-type 10-year simulation at 3.5-km mesh (about 5 years completed), we have statistically confirmed a successful reproduction of many atmospheric aspects over a wide range of spatio-temporal scales (e.g., mean precipitation and OLR distributions, mid-latitude westerly jets, monsoon, MJO, tropical cyclone intensity, and precipitation diurnal cycle). Meanwhile, we also face several issues for high-resolution climate modeling such as the non-triviality of the appropriate representation of low clouds in K-scale NICAM. In this talk, the above recent progress and challenges in our activity are presented, including the perspectives from process-oriented diagnostics and resolution dependency. A current status of simulations with the ocean-coupled NICAM will also be introduced.