My question is about horizontal diffusion and Viscosity coefficient. As you know TNU2 and TNU4 are same in all test (equal to zero) and VISC2 value has been changed in different tests.my grid cells are about 3300 m and changed to 31 m for child grid. How can recognize the appropriate values for those coefficient? What are they depend on?
Cheers
pooran
TNU2 and VISC2 values
-
pooran
- Posts: 72
- Joined: Fri Jan 31, 2014 5:54 pm
- Location: Khoramshahr Marine Science and Technology Universi
Re: TNU2 and VISC2 values
Dear kate
thanks for your reply.
If it is possible please tell me what are the upper and lower bound for TNU2 and VISC2.
Cheers
pooran
thanks for your reply.
If it is possible please tell me what are the upper and lower bound for TNU2 and VISC2.
Cheers
pooran
Re: TNU2 and VISC2 values
Obviously, the lower bound is zero. The upper bound is turning your ocean into a very thick viscous material. If you set them to zero, you might notice that you develop some 2 dx noise in the velocity fields especially. The main point of the viscosity is to damp that out, so you want the bare minimum to do that job without destroying your solution. You can work out the numbers based on your dx and the scale of the features you are interested in, plus the rate of growth of the numerical noise and how long you want the features of interest to last.
In other words, if your eddies of interest are 5 km across, you need a low viscosity to keep them around for a while. That is consistent with needing fine resolution so that your low viscosity is enough to keep the numerical noise from getting out of hand.
In other words, if your eddies of interest are 5 km across, you need a low viscosity to keep them around for a while. That is consistent with needing fine resolution so that your low viscosity is enough to keep the numerical noise from getting out of hand.
-
arango
- Site Admin
- Posts: 1367
- Joined: Wed Feb 26, 2003 4:41 pm
- Location: DMCS, Rutgers University
- Contact:
Re: TNU2 and VISC2 values
Your question cannot be answered in this forum because pretty much you need to be aware of the full history of Physical Oceanography and Ocean Modeling. There are decades of excellent literature about the subject starting with Henry Stommel, Walter Munk, and others. Any Physical Oceanographer or Ocean Modeler needs to read the book The Evolution of the Physical Oceanography . This book was written in honor to Henry M. Stommel, the leading physical oceanographer of the 20th Century.
We cannot, in this forum, provide you with an education in Physical Oceanography. This has been the subject of may graduate curriculum in Physical Oceanography around the world.
The horizontal viscosity in ROMS is modeled as a deviatory stress tensor. Similarly, the horizontal diffusion is modeled as a Redi tensor. Both tensors model physical process in the ocean. These stresses are real. As a byproduct, the viscosity and diffusion improves numerical stability.
We cannot, in this forum, provide you with an education in Physical Oceanography. This has been the subject of may graduate curriculum in Physical Oceanography around the world.
The horizontal viscosity in ROMS is modeled as a deviatory stress tensor. Similarly, the horizontal diffusion is modeled as a Redi tensor. Both tensors model physical process in the ocean. These stresses are real. As a byproduct, the viscosity and diffusion improves numerical stability.