Using the Moist Static Energy Variance Budget to Evaluate Tropical Cyclones in Climate Models against Reanalyses and Satellite Observations

Abstract Tropical cyclones (TCs) have been investigated in general circulation models (GCMs) as well as reanalyses. However, it is known that GCMs and reanalyses struggle to accurately represent TCs and have biases in the TC frequency and intensity distributions. By employing the column-integrated moist static energy (MSE) variance budget as a process-oriented diagnostic, we can analyze the model representation of physical processes that help or hinder the development of TCs. Specifically, we analyze the radiative and surface flux feedback terms within the MSE variance budget across 19 GCM simulations and compare their representation to five reanalysis datasets. Reanalyses are only partially constrained by observations; therefore, we newly introduce satellite-based estimates of the cloud-radiative feedbacks for comparison. Binning and compositing of TC snapshots by two different intensity metrics allows for comparison across the GCMs and reanalyses of the feedbacks at a given intensity, as well as an examination of how the feedbacks vary with storm intensity. Across all models, the longwave feedback positively contributes to TC development and is driven by cloud effects, but is underestimated at weaker intensities compared to satellite observations. Models differ regarding the role of the shortwave feedback, though it is generally smaller than the longwave. Regarding the intermodel spread in surface flux at a given intensity, models with coarse grid spacing tend to have weaker surface flux feedbacks than those that are more finely gridded. These results can help us understand how differences in model representation of physical processes lead to model differences in TC intensity.