reasonable value for maximum Haney number?
reasonable value for maximum Haney number?
The suggested maximum rvalue, or Beckmann and Haidvogel number, has been previously discussed on this forum, but what about the Haney number? Shchepetkin and McWilliams (2003) test with a value just under 4; is that considered a reasonable maximum?
Beckmann-Haidvogel "rx0" number is relevant only because it controls Haney number "rx1" (under all other conditions -- number of vertical levels and stretching -- other kept the same, the two are proportional to each other), and also because you know Haney number only after you setup your vertical grid, while in practice you have to make decisions about topography much earlier in the process of setting up your configuration.
Coarser vertical resolution and stretching more concentrated on top usually mean that you can tolerate steeper topography.
"rx1" up to 3 is considered "safe and conservative"; less than that value usually means that you smooth your topography too much in a realistic application; In practice we often run model having maximum values up to 5 or 6; above that you have to use more caution. Values exceeding 8...10 are considered "insane".
Note that when rx1 = 1 (meaning exactly 1), the hydrostatic error of POM density Jacobian scheme vanishes identically.
Also note that even "rx1" is not the whole story: remember than no or weak and nearly uniform stratification means noor small error regardless of steepness of topography (hence you do not have to smooth too much very shallow areas in marginal seas because they are typically too shallow to be strongly stratified). So the above recommendations are to be considered as the outcome of practical experience, and not mathematically well-defined thresholds.
One can always test it by running a rest stratification test using a representative vertical profile and observe levels of artificial currents. Then decide whether they are acceptable. Representative means to pick a profile at a single point, and do not average it over horizontal directions: averaging usually results in a much smoother vertical profile which is not representative in sense of errors it generates.
Coarser vertical resolution and stretching more concentrated on top usually mean that you can tolerate steeper topography.
"rx1" up to 3 is considered "safe and conservative"; less than that value usually means that you smooth your topography too much in a realistic application; In practice we often run model having maximum values up to 5 or 6; above that you have to use more caution. Values exceeding 8...10 are considered "insane".
Note that when rx1 = 1 (meaning exactly 1), the hydrostatic error of POM density Jacobian scheme vanishes identically.
Also note that even "rx1" is not the whole story: remember than no or weak and nearly uniform stratification means noor small error regardless of steepness of topography (hence you do not have to smooth too much very shallow areas in marginal seas because they are typically too shallow to be strongly stratified). So the above recommendations are to be considered as the outcome of practical experience, and not mathematically well-defined thresholds.
One can always test it by running a rest stratification test using a representative vertical profile and observe levels of artificial currents. Then decide whether they are acceptable. Representative means to pick a profile at a single point, and do not average it over horizontal directions: averaging usually results in a much smoother vertical profile which is not representative in sense of errors it generates.