Evaluation of gravity wave parameterization schemes in a climate model using high-resolution ICON and IFS simulations

Expanding upon our previous work1, we extend the evaluation of gravity wave parameterization schemes in the Atmospheric Component of the IPSL Climate Model (LMDZ6A) by incorporating comparisons with high-resolution datasets from the ICOsahedral Nonhydrostatic Weather and Climate Model (ICON) and the Integrated Forecasting System (IFS). The ICON dataset corresponds to ~ 5 km horizontal resolution simulations for spring 2020, coarse-grained to a ~ 100 km grid (1°). The IFS dataset corresponds to 1 km horizontal resolution simulations for winter 2018, coarse-grained to a T42 grid (~2.8°). In both models, we assume that at each time and place in the stratosphere, the momentum fluxes due to the disturbances that are filtered out during coarse graining are due to subgrid-scale gravity waves. The parameterizations have been then run offline using ICON and IFS coarse-grained meteorological fields to predict these subgrid-scale gravity wave momentum fluxes. The comparison shows that the parameterizations have some skills in predicting the geographical distribution of the simulated fluxes in different regions. More specifically, the gravity wave momentum fluxes due to the orographic and convective gravity waves are reasonably well predicted in the mountainous and tropical regions, respectively. The results are more contrasted concerning the gravity waves generated within fronts. Aloft the storm tracks the parameterized gravity wave momentum fluxes are larger than the ICON gravity wave fluxes and smaller than the IFS gravity wave fluxes. This challenges the dynamics at work in these models during geostrophic adjustment, suggesting that some high-resolution models potentially produce more gravity wave fluxes than are needed in GCMs to simulate the right climate. These results also highlight the importance of considering multiple high-resolution datasets to understand gravity wave characteristics better and tune their parameterizations more effectively.Using insights from these comparisons, we vary the parameters in the schemes to improve the fit with the high-resolution simulations and test impacts in online runs done with the  LMDZ6A climate model. Our results illustrate how high-resolution model datasets can improve gravity wave parameterizations in climate models.  1Toghraei, I., Lott, F., Köhler, L., Stephan, C., and Alexander, J.: Comparison between the gravity wave stress parameterized in a climate model and simulated by the high-resolution non-hydrostatic global model ICON, EGU General Assembly 2024, Vienna, Austria, 14–19 Apr 2024, EGU24-5181, https://doi.org/10.5194/egusphere-egu24-5181, 2024.