2:00 PM - 2:15 PM
[AAS05-02] Implementation and Verification of Super-Parameterized MIROC with Blockwise Coupling
★Invited Papers
Keywords:Global climate model, Super-parameterization, Effective resolution
Atmospheric circulation has a wide range of spatial and temporal scales with mutual interaction among them, complicating the numerical simulations with a high accuracy. In particular, state-of-the-art global climate models (GCMs) cannot resolve small-scale cumulus convections, causing a failure to reproduce sufficient amplitudes of convectively coupled large-scale atmospheric disturbances including equatorial Kelvin waves (EKWs).
Super-parameterization (SP) is a low-cost technique to simulate multiscale interactions explicitly. In usual super-parameterized GCMs (SP-GCMs), a cloud-resolving model (CRM) that calculates small-scale processes including cumulus convections is embedded in each GCM grid, so that conventional cumulus parameterization is no longer used. A scale separation is assumed between GCM grids and CRM domains in SP, which becomes questionable as the horizontal GCM resolution goes higher. To cope with this difficuly, a technique called blockwise coupling is devised and verified in this study. Under the blockwise coupling, a horizontal average of multiple GCM columns, instead of one, is coupled to one domain of the CRM component. This enables SP-GCMs to bypass the problem of shrinking the CRM domain in high-resolution GCMs and furthermore it can drastically reduce the computational cost.
A blockwise-coupled SP-GCM, named SP-MIROC, is developed by coupling the climate model MIROC6 to the cloud-resolving model SCALE-RM v5.3.6. GCM columns are horizontally bundled to 4x4-sized blocks, each of which is coupled to one CRM domain. Contrary to conventional SP-GCMs, the planetary boundary layer scheme is retained in the GCM column because spatial averaging operated in the blockwise coupling filters out friction to small-scale rotational winds and destabilizes simulations.
The blockwise-coupled SP-MIROC successfully reproduced horizontal patterns and frequency distributions of precipitation and realistic amplitudes of EKWs, which were too weak in the standard MIROC (CTL-MIROC) without the SP, while a bias of excessive ice clouds is identified in the tropical upper troposphere. Comparison of snapshots and power spectra between the blockwise-coupled SP-MIROC and the conventional, non-blockwise SP-MIROC indicated that the effective resolution of dynamic variables is not degraded by the blockwise technique. Rather, power spectra of dynamic variables in the blockwise-coupled SP-MIROC are found to be more realistic than those in the conventional SP-MIROC. The 4x4-bundling could offer the best match of resolutions between the dynamics and the physics because the effective resolution of the dynamics is lower than the nominal grid spacing.
Super-parameterization (SP) is a low-cost technique to simulate multiscale interactions explicitly. In usual super-parameterized GCMs (SP-GCMs), a cloud-resolving model (CRM) that calculates small-scale processes including cumulus convections is embedded in each GCM grid, so that conventional cumulus parameterization is no longer used. A scale separation is assumed between GCM grids and CRM domains in SP, which becomes questionable as the horizontal GCM resolution goes higher. To cope with this difficuly, a technique called blockwise coupling is devised and verified in this study. Under the blockwise coupling, a horizontal average of multiple GCM columns, instead of one, is coupled to one domain of the CRM component. This enables SP-GCMs to bypass the problem of shrinking the CRM domain in high-resolution GCMs and furthermore it can drastically reduce the computational cost.
A blockwise-coupled SP-GCM, named SP-MIROC, is developed by coupling the climate model MIROC6 to the cloud-resolving model SCALE-RM v5.3.6. GCM columns are horizontally bundled to 4x4-sized blocks, each of which is coupled to one CRM domain. Contrary to conventional SP-GCMs, the planetary boundary layer scheme is retained in the GCM column because spatial averaging operated in the blockwise coupling filters out friction to small-scale rotational winds and destabilizes simulations.
The blockwise-coupled SP-MIROC successfully reproduced horizontal patterns and frequency distributions of precipitation and realistic amplitudes of EKWs, which were too weak in the standard MIROC (CTL-MIROC) without the SP, while a bias of excessive ice clouds is identified in the tropical upper troposphere. Comparison of snapshots and power spectra between the blockwise-coupled SP-MIROC and the conventional, non-blockwise SP-MIROC indicated that the effective resolution of dynamic variables is not degraded by the blockwise technique. Rather, power spectra of dynamic variables in the blockwise-coupled SP-MIROC are found to be more realistic than those in the conventional SP-MIROC. The 4x4-bundling could offer the best match of resolutions between the dynamics and the physics because the effective resolution of the dynamics is lower than the nominal grid spacing.