We used several tools to evaluate how well each grid size captured key flow features like swirling eddies, shear layers, and return currents. These tools included turbulence energy spectra, statistical errors, information theory metrics, and wavelet analysis to study the flow in both time and space.

We found that medium and fine grid sizes were good enough to capture important turbulence features, while coarser grids showed noticeable information loss. Interestingly, beyond a certain grid resolution, making the model even finer didn’t add much new information, but increased computational costs.