Project No: 16309025

Title: Developing a Scale-Adaptive Cumulus Parameterization for Simulating Tropical Convection Across Scales

Principal Investigator: Prof. Xiaoming SHI

Abstract:

With advancements in computing power, regional numerical weather prediction (NWP) and climate simulations have increasingly adopted kilometer-scale resolutions. Some modeling centers have also experimented with global atmospheric simulations at these high resolutions. While such enhancements have improved our ability to resolve finer details of weather systems, simulating convective systems at these scales remains challenging. Kilometer-scale resolutions fall within the gray zone of convection parameterization. Traditional cumulus convection parameterization schemes rely on assumptions that fail at these scales; therefore, directly applying them can reduce simulation fidelity. Conversely, simply turning off convection parameterization leads to issues in the initiation, organization, and intensity of simulated convective systems. Efforts have been made to modify conventional cumulus convection parameterizations to make them scale-aware so that as resolution is refined, their effects are gradually diminished. This strategy is beneficial in convection-permitting simulations and variable-resolution simulations that include both coarse and kilometer-scale resolutions. Nevertheless, recent work by Shi and Wang (2022) and Wang et al. (2024) suggests that in the gray zone, combining a mass-flux cumulus scheme with an effective turbulence model for horizontal mixing can enhance the simulation of tropical cyclones at kilometer-scale resolutions. This project aims to build on our previous work by developing a three-dimensional, scale-adaptive convection parameterization for gray-zone and variable-resolution simulations of convective weather systems. It will utilize the new Tiedtke scheme to parameterize mass fluxes in convection and the Reconstruction and Nonlinear Anisotropy (RNA) turbulence model for partially resolved turbulence fluxes in and around convection. Termed 'scale-adaptive,' the new scheme adjusts based not only on the grid spacing in a simulation but also on the life cycle stages of convection. At the same convection-permitting resolution, convection that cannot be fully resolved should be primarily represented with the mass-flux scheme, while mature convection that can be sufficiently resolved requires only the turbulence scheme to model subgrid-scale fluxes. We will design, test, and optimize such a scale-adaptive convection scheme using convection-permitting simulations with the Cloud Model 1 (CM1) for various cases of organized convection. Variable-resolution simulations will also be employed to test the new scheme's capability to enable high-fidelity convection simulations across scales.