Parameterization Errors in Numerical Simulations of Stably-Stratified Atmospheric Boundary Layers Using CASES-99 Observations

Northwest Research Associates, Incorporated, Seattle WA

Investigators

Abstract

Understanding the dynamics of the atmospheric boundary layer is crucial to better understanding of both weather and climate. A numerical study of the evolution of a stably stratified atmospheric boundary layer (ABL) will be carried out using a large-eddy simulation (LES) technique and a mesoscale model. This study will combine numerical simulations and observations from a recent field program - CASES-99. Numerical simulations will complement observations in providing more detailed information about stable boundary layer (SBL) processes: turbulence transport and mixing of momentum, heat, and moisture. The goals of the research are: 1) to conduct long-term, high-resolution simulations of canonical, strongly stratified nocturnal boundary layers (NBL); 2) to verify the results of these simulations using an extensive observational data base; and 3) to develop and test more accurate parameterizations of SBLs for large-scale simulations. Underlying deficiencies in current atmospheric turbulence and surface layer parameterizations result in their failure to represent accurately global intermittence under strong stably-stratified conditions. The hypothesis that the failures of parameterization arise from their inability to account for nonlinear effects such as breaking gravity waves and Kelvin-Helmholtz billows will be tested. Parallel LES and mesoscale codes will be used on a massively parallel processor to resolve all the dynamically relevant scales of motion from several meters (small inertial range eddies) to several tens of kilometers (energy containing eddies). Special attention will be given to represent accurately interactions between gravity waves and shear associated with low-level jets and conditions under which global intermittence occurs. By resolving a large fraction of turbulence eddies as well as gravity waves it will be possible to reduce uncertainties and biases introduced by turbulence parameterizations in mesoscale models.