Process-level investigation of scale sensitivity in simulating tropical convection by the ACME V1 model
Authors
Wuyin Lin — Brookhaven National Laboratory
Shaocheng Xie — Lawrence Livermore National Laboratory
Satoshi Endo — Brookhaven National Laboratory
Andrew M. Vogelmann — Brookhaven National Laboratory
Jean-Christophe Golaz — Lawrence Livermore National Laboratory
Yun Qian — Pacific Northwest National Laboratory
Minghua Zhang — Stony Brook University
Category
Convective clouds, including aerosol interactions
Description
As part of the Regional Refined Mesh (RRM) CMDV project (CMDV-RRM), this work aims to develop a framework for evaluating the scale sensitivity of the ACME model in simulating tropical convection. The global ACMEv1 model to some extent has to be tuned separately for high- and low-resolution configurations. This is both a necessity given the existing physical parameterizations are not readily scale aware, and an inconvenience for evolving model development toward increasingly higher resolution. Better understanding of scale sensitivity will aid in the tuning process of the current model, and the metrics established for process-level analysis will benefit future scale-aware parameterization development in quantifying the simulation skills across the scale. The ARM TWP-ICE case is used to study the scale sensitivity in simulating tropical convection. Baseline simulations using reasonably tuned ACMEv1 high- and low-resolution configurations are first analyzed to describe the overall scale sensitivity in this tropical convective regime. The single-column model (SCM) configuration of the ACMEv1 will then be extensively used to investigate the scale sensitivity at process level, taking into account the uncertainty in physical parameter space, through perturbations from the respective baseline high- and low-resolution configurations. The weak temperature gradient approximation approach will be used to account for the large-scale feedback in SCM simulations. Cloud-scale dynamical properties for process-level analysis will initially use observationally constrained CRM simulations driven by the ARM forcing product. The ARM radar-based retrievals will be incorporated as the study progresses. The framework established in this work will be a cornerstone for more systematic analysis of scale sensitivity of the ACME model using RRM configurations with varying refinement grids.
