Mesoscale convective systems (MCSs) are responsible for majority of the total rainfall and latent heating in the tropics. However, it is often difficult to use observations or global reanalyses to examine the detailed dynamic and thermodynamic characteristics of tropical MCSs, as well as their multiscale interactions with the large-scale circulations. This difficulty is due to multiple factors. First, the tropical atmosphere is only sparsely observed by ground-based observation networks. Second, while satellite observations are frequently available at higher spatiotemporal resolutions, they cannot provide information on all the required meteorological fields. Finally, current global reanalysis datasets do not have the ability to explicitly resolve tropical MCSs.

In this talk, we will introduce a new convection-permitting tropical MCS reanalysis dataset. This new dataset generated using advanced ensemble-based data assimilation techniques and covers majority of the tropical Indian Ocean, tropical continental Asia, and the Maritime Continent.

There are two major advantages that this new high-resolution reanalysis dataset has over current global reanalyses. First, this new high-resolution reanalysis dataset can explicitly capture individual tropical MCSs. Second, this new dataset provides detailed and observation-constrained three-dimensional dynamic, thermodynamic, and microphysical information about tropical MCSs. These advantages over current global reanalysis datasets are achieved through a combination of convection-permitting regional model, high-resolution geostationary satellite all-sky infrared observations, satellite-observed atmospheric motion vectors (AMVs), and in-situ observations.

In order to establish the quality of this new high-resolution convection-permitting reanalysis, we compared the performance of our reanalysis against the state-of-the-art European Center for Medium-Range Weather Forecasts Reanalysis version 5 (ERA5). Our reanalysis considerably outperforms the ERA5 in terms of the cloud fields and rainfall. Furthermore, hindcasts from our reanalysis has significantly better cloud fields and rainfall than that from the ERA5, and these advantages can last for more than 8 hours of forecast lead time. In terms of the large-scale environment, our reanalysis has a performance similar to that of the ERA5. Process-orientated analyses are further performed to track each MCS in the high-resolution reanalysis, and to compare the analyzed MCSs’ characteristics with independent satellite observations/retrievals. The results suggest that this new convection-permitting tropical MCS reanalysis can be a useful high-resolution gridded data source for studying the dynamics and thermodynamics, and multiscale interactions of tropical MCSs.