The simulations are conducted with a finite-difference numerical model (NH3D) based on the non-hydrostatic equations of motion in sigma (normalised pressure) coordinates and the idealised mountains considered have a conical shape (with the sharp summit removed for numerical stability reasons) with slopes between 0.25 and 1. The model is configured so that the computational domain width and height (relative to the mountain) are identical to those in a set of laboratory experiments conducted in a stratified towing tank at the Environmental Flow Research Centre (EnFlo), University of Surrey, UK. The simulations are given an initial perturbation (in the potential temperature field) in order to promote asymmetry in the mountain wake and vortex shedding. By comparing results from simulations with and without the perturbation it is clear that the vortex shedding makes a significant contribution (up to 50 %) to the total drag when F < 0.3. This agrees well with the laboratory results. The frequency of the vortex shedding, f, in the numerical simulations is found to be independent of height and the non-dimensional shedding frequency, S=fL/U, where L is the cone base width compares well with the laboratory measurements.
Analysis of the vertical vorticity field in the wake reveals that, although the vortex shedding frequency is independent of height, the phase of the vertical vorticity changes with height and, depending on the mountain slope, the vertical vorticity phase lines can show a significant downwind tilt. These results are compared with laboratory results and an explanation for the phase tilt is sought.
![](images/flag.png)
![](images/blue_ribbon.png)