A129-01
Development, Implementation and Process-Based Evaluation of a New Unified Boundary Layer and Convection Parameterization in Climate Models: The EDMF Approach
Development, Implementation and Process-Based Evaluation of a New Unified Boundary Layer and Convection Parameterization in Climate Models: The EDMF Approach
Friday, 11 December 2020: 10:30
Virtual
Abstract:
The overarching goal of this work is to reduce key biases related to boundary layer clouds and deep convection in climate models, by implementing, and evaluating, a new unified boundary layer and convection parameterization based on the multi-plume Eddy-Diffusivity/Mass-Flux (EDMF) approach. This is a turbulence and convection parameterization that can be considered fully unified, since it is able to represent convective processes from boundary layer convection (dry and cloudy) to deep moist convection with one single parameterization. This work is focused on the boundary layer and the transition to deep convection – in particular, (i) the spatial transition over the oceans from stratocumulus to cumulus (shallow convection) topped boundary layers and to deep convection, and (ii) the temporal transition (diurnal cycle) over land from dry convection, to shallow convection and to deep precipitating convection. A critical component of this work has been the development of hierarchical process-based model evaluation tools ranging from single column model (SCM) comparisons with Large-Eddy Simulations (LES) to SCM (forced by re-analysis) evaluation versus satellite and in-situ observations over long periods of time, and large regions, and finally to parameterization evaluation within three-dimensional climate models (in aquaplanet, atmosphere-only and coupled modes). In this presentation, we will discuss the implementations of EDMF in the National Center for Atmospheric Research (NCAR), the Geophysical Fluid Dynamics Laboratory (GFDL) and the Department of Energy (DOE) climate models and the process-based evaluation of EDMF using the tools mentioned above.