Tropical convection is an important component of the general circulation due to its role in driving large scale circulations which redistribute energy received at the equator to higher latitudes. The behavior of these large scale circulations is sensitive to the vertical distribution of diabatic heating produced by mesoscale precipitation system, which are driven by precipitation formation processes. Convection that occurs in the western Pacific warm pool plays a particularly important role in driving large scale circulations such as the Walker circulation and the inter-annual El Nino-Southern Oscillation (ENSO), which is influenced by the 30-60 day intraseasonal oscillation (ISO). Budget studies of diabatic heating profiles reveal two basic modes of heating: one with positive heating throughout the depth of the troposphere associated with convective precipitation, and another with positive heating in the upper troposphere and negative heating (cooling) in the lower troposphere associated with stratiform precipitation. Because stratiform heating profile shapes vary little from region to region and case to case, variations in total (convective plus mesoscale) heating profile shapes are thought to be linked to variations in the shape of convective heating profiles (Houze, 1989). However, long-term observations of convective vertical structures in the tropics-particularly the western Pacific warm pool--are either non-existent or limited to a few locations. As a component of the recently completed Tropical Ocean Global Atmosphere (TOGA) Coupled Ocean-Atmosphere Response Experiment (COARE), the MIT 5-cm Doppler radar, mounted aboard the NOAA Research Vessel John V. Vickers, was deployed to a fixed location within the warm pool region for three approximately 30-day periods and continuously monitored the three-dimensional structure of precipitating systems. In this study, radar data were partitioned into convective and stratiform components in order to address two main goals: 1) to determine the characteristic distributions of convective vertical structure during COARE and their variability over time, and 2) to relate observed variations of convective vertical structure to larger scale environmental variables. Distributions of convective feature heights and 30 dBZ contour heights (an indicator of convective vigor) reveal that convective activity is modulated by the phase of the ISO as well as by intrusions of low-level dry subtropical air. However, at least some deep convection was nearly always present. The most intense convection--as observed by convective feature reflectivity profiles--occurred in environments with the greatest thermal buoyancy (the vertical distribution of CAPE) experienced by a parcel lifted from the mixed layer. However, these periods were also characterized by strong inhibitors to convective development which limited intense convective activity to just a few days. Convective heating profiles-computed from a combination of budget-derived total beating, radar-derived rainfall characteristics, and ideal and computed radiative heating profiles-varied in a manner consistent with the variations in vertical reflectivity and thermal buoyancy profiles. Namely, less vertically intense convection was associated with a lower altitude convective heating maximum than was observed for those days with convective reflectivity profiles indicative of more vigorous convection. Possibilities for future research include a rigorous comparison of convective characteristics between the COARE and GATE regions, modeling studies of the influence of low-level dry air on convective activity and its relation to tropospheric drying and testing a refinement of passive microwave rainfall retrieval algorithms.