Richard Neale (NCAR) November 4, 2010 Understanding and Modeling the Tropical Atmosphere: Phenomena, Processes and Resolution Techniques for representing convective processes in global models are in a transitional phase. Existing assumptions for parameterizing convection are increasingly inadequate due to ever finer horizontal grids and increasing demands on their performance, but a transition to global cloud-resolving models is still severely resource limited. In this interim period climate models will need to provide useful estimates of extreme weather statistics against a potentially rapidly changing underlying climate. Convective processes lie at the heart of an adequate model simulation of tropical variability and many of the key convective characteristics such as sensitivity to humidity and the strength of organization are common across scales. A poor representation of entrainment leads to over-sensitivity to surface forcing and under-sensitivity to tropospheric humidity - detriments to the diurnal cycle, MJO and ENSO simulations. Coherent organization is largely absent at the sub-grid scale which is detrimental to understanding convective system propagation and simulating the correct prevalence of extreme precipitation events. This talk will outline existing and ongoing modeling efforts to improve the representation of these important aspects of convective processes, detail their grounding in observations and demonstrate their application and effects across scales, from the diurnal cycle through ENSO.