Differences among three types of tropical deep convective clusters observed from A-Train satellites

Deep convective clusters (DCCs) constitute a complex regime with a variety of types of cloud parameterization in multiple global climate models (GCMs). This study develops a method to identify tropical DCCs and presents regional differences in cloud occurrence, surface precipitation and a vertical survey of ice water content among separated DCCs, connected DCCs and mesoscale convective systems (MCSs) determined by collocating data from MODIS, AMSR-E, CPR and CALIOP instruments on board A-Train satellites. The results reveal that separated DCCs and MCSs occur frequently over East Asia, while connected DCCs are more common over the warm pool of the West Pacific in June-August and over South America in December-February. Higher surface convective precipitation rates occur more frequently in MCSs than in separated and connected DCCs in non-monsoon areas. However, connected DCCs have a large ice water content (IWC) at 10-15 km similar to that of MCSs, and these clusters are more intense than those in separated DCCs. Remarkably large anvil-stratiform clouds connected with multiple deep convective cores in connected DCCs can be easily confused with MCSs if the cloud top brightness temperature in the horizontal view is the only parameter being considered. Combined observational data of surface precipitation of DCCs will be helpful for distinguishing connected DCCs and MCSs.