A096-0005
First results from a convection-permitting pan-European fully coupled TSMP simulation

Thursday, 10 December 2020
Poster
Klaus Goergen1, Alexandre Belleflamme1, Abouzar Ghasemi2, Carl Hartick1, Bibi S Naz1, Liuba Poshyvailo1, Wendy Sharples3, Niklas Wagner1 and Stefan J Kollet1, (1)Forschungszentrum Jülich GmbH, Institute of Bio- and Geosciences (Agrosphere, IBG-3), Jülich, Germany, (2)Forschungszentrum Jülich GmbH, Jülich Supercomputing Centre, Jülich, Germany, (3)Bureau of Meteorology, Melbourne, Australia
Abstract:
Convection-permitting regional climate models show added value over coarser resolution simulations, e.g., in the reproduction of precipitation features or dynamical processes, and they also represent small-scale surface heterogeneities more accurately. Regional multi-physics coupled Earth system models, which link the interacting compartments of the geo-ecosystem through mass, energy, and momentum transfers, allow for a better representation, e.g., of the terrestrial water cycle and land-atmosphere coupling processes. In this study, we combine both of these aspects in a convection-permitting (3km) continental-scale Terrestrial Systems Modelling Platform (TSMP, https://www.terrsysmp.org) simulation in a configuration with the atmospheric model COSMO v5.01 in climate mode, the Community Land Model (CLM) v3.5, and the ParFlow v3.2 hydrologic model, coupled through OASIS3-MCT. In an evaluation run setup, TSMP is driven with a single nesting step at the lateral boundaries and the sea surface by ECMWF ERA5 reanalysis; the model domain is based on the 12km pan-European EURO-CORDEX grid, but with an increased 3km convection-permitting resolution. The necessary hydrological model subsurface spinup comes from a previous standalone ParFlow experiment with refined land and subsurface parameters. Computational and data challenges with such simulations are large; the overall simulation is run in a MPI-only CPU-only configuration; the big data large volume model output is dealt with by the novel Helmholtz Analytics Toolkit and parallel postprocessing and diagnostics tools. Based on a few years of hindcast simulations, we investigate, across multiple spatial scales, the impacts of the 3D soil- and groundwater representation, linked in a continuity approach with the overland flow in ParFlow, on the (terrestrial) water budget in the context of land-atmosphere coupling. This is interesting as convection-permitting simulations, with a more simplified soil moisture and groundwater treatment, show tendencies towards a stronger coupling, moving towards more moisture-limited coupling regimes, as opposed to coarser resolved simulations, due to more small-scale precipitation patterns, less stratiform clouds, higher global radiation, in combination with a higher surface runoff and less infiltration.