Exploring a Variable-Resolution Modeling Approach within a Global Nonhydrostatic Dynamical Core

High-resolution numerical weather and climate models have a great advantage in both prediction and simulation for their ability to resolve small-scale systems, but suffer from expensive computational cost. The aim of this study is to explore a cost-effective variable-resolution modeling approach within a newly developed global nonhydrostatic dynamical core based on an unstructured mesh. We provide a size-controllable formulation of hierarchical refinement mode by an adapted density function for more realistic high-resolution simulations. The dynamical core is tested regarding both dry and moist atmosphere to evaluate variable-resolution simulations against quasi-uniform ones. In baroclinic wave tests, the variable-resolution model, which owns much less grid points, captures a comparable fine-scale fluid structure with the high-resolution quasi-uniform one in the refinement region. In the coarse region, the result of the variable-resolution simulation matches that of the low-resolution quasi-uniform one, which contributes to smaller global errors of the variable-resolution simulation. A series of sensitivity tests regarding parameters of the hierarchical refinement mode validate the high stability of the variable-resolution model to preserve the intensity and vertical structure of tropical cyclones moving through the transition zone. The variable-resolution modeling lays a strong foundation for potential improvement of regional high-resolution simulations.