• Advection Schemes in LES of Atmospheric Boundary Layer
1. The influence of WENO schemes on resolved turbulence
Using values at multiple grid points to compute fluxes of a variable may generate spurious oscillations when sharp gradients are present. Advection schemes designed to avoid such spurious oscillations (e.g., WENO schemes), however, may also suppress the development of physically realistic instabilities. This work explores the influence of advection schemes on numerical simulations resolving turbulent motions, which are produced through instability development. The analysis focuses on the atmospheric surface layer, where the production of turbulence is highly active. Major findings include recommendations on the use of advection schemes and a new concept to reproduce turbulent motions inadequately resolved by simulation grids.
Publication: Wang, A., Y. Pan, and P. M. Markowski (2021): The Influence of WENO Schemes on Large-Eddy Simulations of a Neutral Atmospheric Boundary Layer. J. Atmos. Sci., 78, 3613-3628, doi: 10.1175/JAS-D-21-0033.1.
• Wall Model (Lower Boundary Condition) in LES of Tornado
1. Turbulence memory
We model the effect of turbulence memory in the wall model and see its influence on an idealized tornado. Turbulence memory means the lag of Reynolds stress to velocity gradient, which is neglected in the traditional equilibrium wall model (semi-slip lower boundary condition) and most eddy-viscosity-type SGS model. When the time scale of an air parcel passing through a curved trajectory is comparable to turbulence’s integral time scale, the remembered previous surface stress may enhance the convergence of angular momentum and strengthen a tornado.
Publication: Wang, A., Y. Pan, and P. M. Markowski (2020): The influence of turbulence memory on idealized tornado simulations. Mon. Wea. Rev., 148, 4875-4892, doi: 10.1175/MWR-D-20-0031.1.
2. Two-layer model
A traditional wall model in LES assumes an instantaneous equilibrium following Monin-Obukhov Similarity Theory, which assumes horizontal homogeneity and quasi-steady state. Tornado is a highly-curved and unsteady flow, thus a more suitable wall model is needed. Two-layer model (Balaras and Benocci, 1996) applies a set of simplified RANS-type momentum equations under the first grid level, allowing the advection and pressure gradient force to break such instantaneous equilibrium, thus yields a more realistic surface drag. We are testing this wall model on idealized atmospheric boundary layer and idealized tornado genesis.
Publication: Wang, A., Y. Pan, George Bryan, and P. M. Markowski (under review): Modeling Near-Surface Turbulence in Large-Eddy Simulations of a Tornado: An Application of Two-Layer Models.
• Field Measurement of Boundary Layer Turbulence
I participated in Dr. Ying Pan’s project of tower measurement of boundary-layer turbulence in Utqiagvik, AK. The research aims to understand the turbulence in a very stable boundary layer like that in the Arctic. In addition to collecting and maintaining the instruments on the tower, I also help collecting the ozone and aerosol data for CHACHA (CHemistry in the Arctic: Clouds, Halogens, and Aerosols) project.