angle, wind, ocean current

[Joeailvyou]

Dear arango, Kate, and wilkin,
I have looked into many discussions about the grid angle and the rotation of wind and ocean current velocity, but got more confused !
1. When I prepare the wind forcing, should I rotate the wind direction from geometrical eastward/northward to grid xi/eta in my forcing .nc file ?
2. When I prepare the ocean current velocity initial, boundary and climatology .nc files, should I rotate their direction from geometrical eastward/northward to grid xi/eta ?
3. I found that if I have defined #CURVGRID in the preprocessor, the /ROMS/Nonlinear/set_data.F rotate the wind again ! So, dose that mean, if I have defined #CURVGRID, I must not rotate the wind direction in my forcing .nc file ?
4. I also found that the /ROMS/Nonlinear/set_data.F doesn't rotate the ocean current velocity, so, there is no relationship between rotation of ocean current velocity and #CURVGRID ?
5. If CURVGRID and wind forcing file ratation are repeated in question 3. There is pretty large angle between xi/eta and east/north in my grid, is it reasonable to undefine CURVGRID and use a rotated wind forcing ?
6. Is there relation among these three: A. the rotation of the wind and ocean current; B #CURVGRID in the preprocessor; C spherical/Cartesian in the grid file ?
7. Dose the sitution change with the version of ROMS ?

Looking forward to hearing from you soon.

Wish you good heath and happiness !
Joe Smith

[Joeailvyou]

I have interpolated forcing, initial, boundary and climatological file into roms grid. The key point is whether rotate wind and ocean current velocity.

[wilkin]

1. When I prepare the wind forcing, should I rotate the wind direction from geometrical eastward/northward to grid xi/eta in my forcing .nc file ?

If you are providing Uwind,Vwind in a netcdf file and having ROMS interpolate from the Uwind,Vwind lon,lat coordinates then ROMS will apply the angle rotation in the regridding step. You will receive confirmation that a regrid is being done in the standard output, e.g.:

GET_2DFLD_NF90 - surface u-wind component, 2007-01-01 21:00:00.00
(Grid=01, Rec=8, Index=1, File: Uwind_narr_NENA_2007.nc)
(Tmin= 365.0000 Tmax= 729.8750) t = 365.8750
(Min = -5.36645855E+00 Max = 7.58490888E+00) regrid = T

If you give wind forcing as stresses sustr,svstr on the staggered grid with dimensions that match u,v, then ROMS will not regrid. stdout should confirm regrid = F.

2. When I prepare the ocean current velocity initial, boundary and climatology .nc files, should I rotate their direction from geometrical eastward/northward to grid xi/eta ?

No, initial, boundary and climatology files are all in ROMS space and coordinates. The file formats for initial and climatology are exactly the same (dimensions and coordinates) as ROMS output history and averages (etc.) files. Inspecting the dimensions, you will see that u and v variables have different dimensions corresponding to their respective staggered Arakawa-C grid positions (i.e. they are not simply components of a total vector defined at the model rho points).

The only place that rotation comes in into water velocities is on output if you request east,north velocities

Code: Select all

Hout(idu3dE) == T       ! u_eastward         3D U-eastward  at RHO-points
Hout(idv3dN) == T       ! v_northward        3D V-northward at RHO-points

This will rotate staggered ROMS u,v to earth coordinates and report them on the rho points grid.

3. I found that if I have defined #CURVGRID in the preprocessor, the /ROMS/Nonlinear/set_data.F rotate the wind again ! So, dose that mean, if I have defined #CURVGRID, I must not rotate the wind direction in my forcing .nc file ?
4. I also found that the /ROMS/Nonlinear/set_data.F doesn't rotate the ocean current velocity, so, there is no relationship between rotation of ocean current velocity and #CURVGRID ?

As I say above, do not rotate Uwind,Vwind. Water velocities are in ROMS coordinate directions AND split out to their _u and _v positions.

5. If CURVGRID and wind forcing file ratation are repeated in question 3. There is pretty large angle between xi/eta and east/north in my grid, is it reasonable to undefine CURVGRID and use a rotated wind forcing ?

The documentation is terse, but such as it is, WikiROMS gives this about CURVGRID: https://www.myroms.org/wiki/CURVGRID

IF dmde or dndx or angle is non-zero, then you should activate this option. The curvature of the grid requires additional metric terms in the momentum equations. The rotation of a grid requires the grid angle to be non-zero. The grid angle is needed for such features as to rotate Uwind and Vwind to the grid (for example).

6. Is there relation among these three: A. the rotation of the wind and ocean current; B #CURVGRID in the preprocessor; C spherical/Cartesian in the grid file ?

All ocean current inputs are in ROMS coordinates. See CURVGRID above. I don't know what happens for non-spherical coordinates.

7. Dose the sitution change with the version of ROMS ?

No.

[zhy202303]

你用的哪一个版本的roms呢？?
2023年推出的新版版似乎有些cpp不需要在h文件中定义了