Abstract
Bias associated with representation of clouds, especially low and high clouds, in large scale atmospheric models remains an unsolved problem. Progress toward alleviating this bias is limited despite the decades of community effort expended on developing and advancing microphysics and turbulence parameterizations. DOE’s Energy Exascale Earth System Model (E3SM), one of the state-of-the-art global models, is no exception for this matter. Recently, several studies have demonstrated that higher vertical resolution results in improved representation of high and low clouds for parameterizations similar to those used in E3SM. This is a relatively unexplored area due to limitations in computational resources. This project will bring improved cloud representation to E3SM by implementing a novel computational framework that uses high vertical resolution at a computationally affordable cost. The Framework for Improvement by Vertical Enhancement (FIVE) has been shown to offer better representation of atmospheric boundary layer clouds with reduced cost. FIVE allocates additional prognostic variables in high vertical resolution and these high resolution prognostic variables are used for computing tendencies for selected physical processes. The computed tendencies from the Vertically Enhanced Physics are then passed to E3SM for predicting E3SM’s prognostic variables. To further reduce the computational cost, we will implement FIVE on an Adaptive Vertical Grid (AVG), which will dynamically adjust vertical resolution depending on the atmospheric state for each grid column. In this presentation, we will discuss a path to E3SM coupled with FIVE-AVG, challenges for developing a computationally efficient FIVE-AVG, and current progress.