...
| Page Properties | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
|
Title
Evaluation of Scale-awareness of dCAPE-based Closure in Zhang-McFarlane Convection Scheme
Authors
Guang J. Zhang, Xu Wang, Scripps Institution of Oceanography, La Jolla, CA
Shaocheng Xie, Lawrence Livermore National Laboratory, Livermore, CA
Abstract
Convective parameterization is one of the major factors responsible for biases in global climate model (GCM) simulations. At spatial scales of ~100 km or larger, there exists a quasi-equilibrium between convection and large-scale forcing. At grey zone scales (e.g., ~10 km or smaller), many important assumptions in convective parameterization break down. Therefore, for high-resolution E3SM development, a scale-aware convection scheme is needed. The existing approaches to making a convective scheme scale-ware is by incorporating a cloud fraction factor as proposed by Arakawa. However, trigger functions and closures for convective parameterization can also affect the scale-awareness of convection schemes. Here we examine a dCAPE-based closure in the Zhang-McFarlane (ZM) scheme using cloud-resolving model simulation output for both tropical and midlatitude convection. To generate GCM grid-scale variables, the CRM output is averaged over subdomains of sizes equivalent to GCM resolutions. The GCM-scale fields are then used to compute dCAPE (CAPE generation by the GCM-grid scale circulation). The relationships between dCAPE so computed and CRM-simulated convection within the corresponding averaging subdomain are examined. It is found that convection (both in terms of precipitation and convective mass flux) is well correlated with dCAPE for GCM resolutions even into grey scale. However, the dependence of convective mass flux on dCAPE becomes nonlinear as the averaging subdomain size decreases to grey zone. The similarities and differences in the relationships between organized and unorganized convection will be presented and discussed.