An event-based method of analyzing the measurements from multiple satellite sensors is presented by using observations of the Tropical Rainfall Measuring Mission (TRMM) precipitation radar (PR), Microwave Imager (TMI), Visible and Infrared Scanner (VIRS), and Lightning Imaging System (LIS). First, the observations from PR, VIRS, TMI, and LIS are temporally and spatially collocated. Then the cloud and precipitation features are defined by grouping contiguous pixels using various criteria, including surface rain, cold infrared, or microwave brightness temperature. The characteristics of measurements from different sensors inside these features are summarized. Then, climatological descriptions of many properties of the identified features are generated. This analysis method condenses the original information of pixel-level measurements into the properties of events, which can greatly increase the efficiency of searching and sorting the observed historical events. Using the TRMM cloud and precipitation feature database, the regional variations of rainfall contribution by features with different size, intensity, and PR reflectivity vertical structure are shown. Above the freezing level, land storms tend to have larger 20-dBZ area and reach higher altitude than is the case for oceanic storms, especially those land storms over central Africa. Horizontal size and the maximum reflectivity of oceanic storms decrease with altitude. For land storms, these intensity measures increase with altitude between 2 km and the freezing level and decrease more slowly with altitude above the freezing level than for ocean storms.

Cold cloud features (CCFs) are defined by grouping six full years of Tropical Rainfall Measuring Mission (TRMM) infrared pixels with brightness temperature at 10.8-μm wavelength (TB11) less than or equal to 210 and 235 K. Then the precipitation radar (PR)-observed precipitation area and reflectivity profiles inside CCFs are summarized and compared with the area and minimum temperature of the CCFs. Comparing the radar with the infrared data, significant regional differences are found, quantified, and used to describe regional differences in selected properties of deep convective systems in the Tropics. Inside 4 million CCFs, 35% (57%) of cold cloud area with TB11 ≤ 235 K (210 K) have rain detected by the PR near the surface. Only ∼1% of the area of TB11 ≤ 210 K have 20 dBZ reaching 14 km. CCFs colder than 210 K occur most frequently over the west Pacific Ocean, but 20-dBZ echoes extending above 10 km in this region are disproportionately rare by comparison to many continental regions. Ratios of PR-detected raining area to area of TB11 ≤ 235 K are higher over central Africa, Argentina, and India than over tropical oceans. After applying these ratios to the climatological Global Precipitation Index (GPI) tropical rainfall estimates, the regional distribution is more consistent with the rainfall retrieval from the PR. This suggests that the discrepancy between GPI- and PR-retrieved rainfall can be partly explained with the nonraining anvil. Categorization of CCFs based on the minimum TB11, size of CCFs, and 20-dBZ heights demonstrates that 20-dBZ echoes above 17 km occur most frequently over land, while the coldest clouds occur most frequently over the west Pacific. The vertical distances between the cloud-top heights determined from TB11 and PR 20-dBZ echo-top heights are smaller over land than over ocean and may be considered as another proxy for convective intensity.

By grouping the contiguous precipitating area detected by the precipitation radar on board the Global Precipitation Mission (GPM) core satellite, snapshots of precipitation systems are summarized as precipitation features (PFs), and their properties are cataloged from 1 year GPM observations. These PFs are categorized by their area and depth and convective intensity based on the 20 and 40 dBZ radar echo tops, respectively. The largest PFs are found mainly over ocean at the mid‐high latitudes, especially over southern ocean. The deepest PFs are mainly over tropical land, the West Pacific Warm Pool, and the Great Plains of the United States and Argentina. The most convectively intense PFs are dominant over land regions, including midlatitude and high latitude. The zonal precipitation contribution from extremely large precipitation systems is greater in midlatitude and high latitude than in the tropics. These extreme precipitation systems are rare but contribute significantly to the global precipitation. It is important to include their impacts in global climate models to correctly describe the global water cycle.

Conventionally, the heaviest rainfall has been linked to the tallest, most intense convective storms. However, the global picture of the linkage between extreme rainfall and convection remains unclear. Here we analyse an 11-year record of spaceborne precipitation radar observations and establish that a relatively small fraction of extreme convective events produces extreme rainfall rates in any region of the tropics and subtropics. Robust differences between extreme rainfall and convective events are found in the rainfall characteristics and environmental conditions, irrespective of region; most extreme rainfall events are characterized by less intense convection with intense radar echoes not extending to extremely high altitudes. Rainfall characteristics and environmental conditions both indicate the importance of warm-rain processes in producing extreme rainfall rates. Our results demonstrate that, even in regions where severe convective storms are representative extreme weather events, the heaviest rainfall events are mostly associated with less intense convection.