Large-Eddy Simulations of the Wind Field over the Vernagtferner Glacier in Austria

Understanding wind flow over complex terrain is critical for accurately modeling surface-atmosphere exchanges and wind field turbulence. This study investigates the dynamics of the atmospheric boundary layer (ABL) over the Vernagtferner Glacier in Austria, using high-resolution large-eddy simulations (LES) under neutrally stratified conditions. The in-house code MGLET is employed to solve the governing equations of motion, incorporating the Coriolis force due to Earth’s rotation as well as a ghost-cell immersed boundary method (GCIBM) to handle complex geometries. A set of simulations was conducted to evaluate the sensitivity of the wind field to variations in domain extent and grid resolution. Special attention was paid to determine how large the numerical domain needs to be to reliably capture the ABL dynamics. Hence, the computational domains were defined as horizontal squares of sizes ranging from 20 km to 60 km, with the glacier at the center, using periodic boundary conditions in the horizontal directions, and the domain top was set between 9 km and 15 km above sea level. For each case, refinement levels were generated to reduce computational effort, with the finest level covering the entire glacier at resolutions ranging from 39 m to 13 m, almost homogeneously in every direction. Additionally, the driving force was a pressure gradient that is in balance with a geostrophic west wind of 10 ms^-1. Results show that smaller domain extents produce less friction between the wind flow and the mountainous terrain, whereas the largest evaluated domain yields higher ABL depth and more intensive turbulence over the glacier. Moreover, refined mesh enhances the magnitude of resolved-scale turbulence and improves the quality of the resolved wind field by reducing numerical oscillations. It is also illustrated how, under the simulated conditions, the majority of the glacier is exposed to elevated turbulence levels, evaluated through turbulence kinetic energy (TKE) and Reynolds stresses, especially near sharp ridges. In addition to wind flow characterization, these findings also provide guidance for the setup of numerical domains and grid resolutions in LES simulations over complex terrain, contributing to improved modeling of wind dynamics and turbulence in mountainous environments.