I'm working on a smaller scale model (dx ~40m) with wind forcing, and am running into issues with waves in the domain. I've tried using a couple of different boundary conditions, but I'm not noticing changes from them. I'm using harmonic mixing, and am using a tnu2 = 0.32 m2 / s , and visc2 = 3.2 m2 / s. I haven't played with the ratios of these numbers, but have varied them.
The grid is 250x245 grid cells, with 40 levels, and a flat bathymetry of 169 m. I also have a sponge layer starting at visc_factor = 10 at the boundaries, and decreasing 1 per grid cell moving inward until it gets to 1. I'm using a baroclinic time step of 2 seconds, with ndtfast = 15
I'm using the doppio history data for forcing and initial conditions. These are both from January 16, 2020
This a cross section of the upward velocities (in m/s) after 40 minutes of simulated time. The vertical axis is just the percent of depth. The slice is taken near the center of the xi axis.
- waveIssues.png (58.51 KiB) Viewed 1178 times
I think the biggest issue is probably with viscosity or diffusivity, but I also don't want to increase it unreasonably, and I don't have a good sense for that. Using the guidance in https://www.myroms.org/wiki/Horizontal_ ... ent_Values , I got a range between 0.16 and 5.7 for nlvisc2, and it sounds like diffusion constants should be between that and 0, with 1/10th being a standard.
