Sensitivity of atmospheric parametric formulations to regional mesh refinement in global climate simulations using CESM-HOMME
Collaborative Institutional Lead
Strategies to better address regional climate problems include the rapidly developing capability of high resolution regional mesh refinement (at least to mesoscale) within global climate model (GCM) simulations. Such an approach will require a scale aware capability of the physical parameterizations in order to provide a more continuous solution as the mesh refines from a globally coarse to a locally fine resolution. The consequences of this requirement for the development of physical parameterizations are likely to be profound. GCM parameter selections are often resolution dependent, both due to unforeseen, unrealistic interactions with resolved dynamical features and due to undesirable sensitivities to time-step and poor formulation inherent to the parameterization's structural deficiencies. Using regional mesh refinement will add the requirement that parameterized processes are well behaved across resolution within a single global domain. Little attention has been paid to the problem of ensuring continuous, convergent solutions across the model domain with and without a varying grid mesh due to variations in physical parameterization parameter settings. The proposed study aims to fill this gap.
The main objectives are to (a) determine the capability of region-specific physical parameterization settings to provide accurate and continuous climate solutions, both local to and remote from the locally refined regions, with and without regional grid refinement, in the Community Earth System Model (CESM) with the High Order Method Modeling Environment (HOMME) dynamical core; (b) to investigate and quantify the inherent uncertainty and stability in climate model solutions as a function of physical parameter settings and consequently develop standard testing procedures, akin to existing dynamical core tests, to identify such solution characteristics; (c) to provide process and observationally- based enhancements to physical parameterizations identified as being most structurally deficient in providing smooth and continuous solution behavior when subject to regional grid refinement and/or physical parameter selections; and (d) to implement and test, both scientifically and computationally, structural improvements to physical parameterizations in refined mesh global simulations.
The major tasks to be performed are: (1) an analysis of the sensitivity of model physical and dynamical parameter values to global resolution and to the regional mesh refinement size, location and degree in idealized, aqua-planet versions of CESM; (2) testing of the most sensitive physics and grid settings to the location of pairs of expected origin and target grids (e.g., West Pacific and North America) in CESM to determine local and global circulation and variability dependencies; (3) the generation of metamodel equivalents of CESM from multi-year perturbed parameter integrations and the optimization of regional parameter settings based on these equivalents; (4) the development, implementation, testing, validation and application of simple scale-aware augmentation to the CESM deep convection parameterization based on stochastic, organization and entrainment grid-scale dependencies.
Outcomes include evaluation tools aimed at providing estimates of robustness in order to guide high resolution model development and configuration decision making, and thus benefiting regional climate change processes and predictions. These benefits target and integrate across existing DOR/BER climate and earth systems modeling programs; specifically Earth System Modeling (ESM) with community models (CESM), Regional and Global Climate Modeling (RGCM) and Atmospheric Systems Research (ASR).