Transpose-AMIP hindcasts of NAWDEX cyclones with the global ICON-NWP and ECHAM6 models

Abstract:

Recent studies have shown that errors in the representation of fast cloud processes have a strong impact on climate projections. Since weather forecasts are also strongly impacted by these processes, this creates a potential to use numerical weather prediction (NWP) perspectives for assessing climate models. To reveal the parametrization errors in models and to understand the origin of climate-scale (long-term) biases in models, we perform short-term weather hindcasts for the North Atlantic Waveguide and Downstream Impact Experiment (NAWDEX) period with the global ICON-NWP model at 40 km resolution and the ECHAM6 model at T127 resolution (about 1 degree resolution). This approach allows us to directly compare the large-scale extratropical circulation and parametrized cloud-relative fields, cloud water content, precipitation and radiative fluxes associated with the cyclone with observations. For each of the two models, we study 17 hindcasts with a length of 5 days initialized from the ECMWF IFS analysis. The simulation setup follows the Transpose-Atmospheric Model Intercomparison Project (short: T-AMIP) approach also applied in previous work. We compare the hindcasts with the ERA5 reanalysis product. We find that cyclone tracks are generally well reproduced in all considered cases by both models, but the deepening is not, with a general delay in the development of the cyclones. The deviations are particularly strong in the initial and mature stages of the cyclones, but better agreement is found for the decaying stage. We believe that this deviation stems from a misrepresentation of diabatic processes within the cyclone, especially along fronts, and we plan to investigate this idea by means of the surface pressure tendency equation.