Cloud and water vapor feedbacks that control the magnitude and timing of global climate change are most sensitive to interactions between tropical diabatic heating anomalies and the general circulation response. On interannual time scales, global climate models routinely underpredict the remote atmospheric response to equatorial sea surface temperature perturbations. Global models also underestimate the stratiform component of convective heating and thus the altitude of the overall diabatic heating peak. These two deficiencies are likely connected, and correcting them is a prerequisite for having confidence in model climate change projections. We propose to conduct several studies, using available decadal length TRMM data products and ancillary data, that will guide cumulus parameterization development and evaluation in these areas. First, we will use TRMM Precipitation Radar (PR) reflectivity profiles and Lightning Imaging Sensor (LIS) lightning occurrence data for land vs. ocean storms identified in a previous cloud object analysis of TRMM Microwave Imager (TMI) data to refine an existing cumulus updraft speed and precipitation-detrainment partitioning parameterization in a global climate model (GCM). Next, we propose to use geostationary satellite data to track identified tropical convective clusters through their entire life cycle. PR and LIS data will be mapped onto these cluster tracks to determine a composite vertical structure, updraft intensity index, surface rain rate, convective-stratiform partitioning, and latent heating profile as a function of life cycle stage for various classes of storms in several regions. Similar statistics derived from a GCM will identify deficiencies in the various stages of simulated storm growth and decay and provide a guide for the development and testing of a mesoscale updraft/downdraft parameterization, and improvement of an existing convective downdraft and rain evaporation scheme, that will simulate realistic heating profiles in different environments. The fidelity of the resulting scheme will be evaluated by conducting global simulations over the entire TRMM record and comparing observed and simulated ENSO anomalies in both rainfall products and associated water vapor and cloud-radiation parameters. We will also conduct an anthropogenic climate change experiment to examine the effect of the resulting parameterization changes on climate sensitivity. Finally, we will investigate the ability of the TMI data products to capture the salient features of the radar-observed composite life cycle structure as a means of evaluating the potential of the proposed Global Precipitation Mission constellation approach for convective cluster evolution research.
