depth

[pooran]

hi
How can compute each sigma level's depth at rho point? By multiplying rho point's depth by value of each level's S_coor Or CS_curve ?is there another way?
Thanks

[kate]

You can see how ROMS is doing it in set_depth.F:

Code: Select all

cff_r=hc(ng)*(SCALARS(ng)%sc_r(k)-SCALARS(ng)%Cs_r(k))
cff1_r=SCALARS(ng)%Cs_r(k)
hwater=h(i,j)
z_r0=cff_r+cff1_r*hwater
z_r(i,j,k)=z_r0+Zt_avg1(i,j)*(1.0_r8+z_r0*hinv)

There is similar code out there for Matlab and Python.

[pooran]

Sorry, but i have used rnt_setdepth.m (which located in RNT/matlib/rnt subdirectory) and depths.m (which is in ROMS/MATLAB/utility) for computing depth at rho points. Results were different
Actually, i my aim is to extract u and v values in specific depth.if you have another recommend i become appreciate to receive it
many thanks in advance

[kate]

Are you asking which of the Matlab routines is correct? The only way to know for sure is to compare with the ROMS code. Perhaps you can use the ROMS output that looks like this to compare to:

Code: Select all

 Vertical S-coordinate System, Grid 01:

 level   S-coord     Cs-curve   Z   at hmin       at hc    half way     at hmax 
    50   0.0000000   0.0000000        0.000       0.000       0.000       0.000
    49  -0.0200000  -0.0000415        0.319      -2.505      -4.728      -5.038
    48  -0.0400000  -0.0001667        0.638      -5.021      -9.686     -10.557
    47  -0.0600000  -0.0003782        0.957      -7.547     -14.882     -16.571
    :
     2  -0.9600000  -0.9012040       14.456    -232.651   -2703.668   -5407.838
     1  -0.9800000  -0.9531675       14.726    -241.646   -2851.406   -5711.170
     0  -1.0000000  -1.0000000       15.000    -250.000   -2985.012   -5985.024

For getting depths at the u,v points, simply average the two nearest rho point depths.

[pooran]

Dear
actually, my first post was about that . I want to know how i can compute each layer depth by ROMS output .
would you please tell me that the correct way to compute depth (by ROMS output) is to multiplying each layer's S_coor (or CS_curve ) with h values which can be gotten from h matrix of grid file?
Best
pooran

[wilkin]

Since it appears you are a Matlab user (you mention using rnt*.m routines in the past) there are Matlab routines on the myroms.org site that compute the z-coordinates.

If you don't want to download the entire set of Matlab routines from the repository, you can browse the code the same way you might browse the ROMS code itself and just grab what you need.

Go to the "trac" source code browser https://www.myroms.org/projects/src/browser and login with the same roms username/password you use to access the User Forum and the myroms.org svn source code. Once logged in you will see a link to the Matlab code tree.

Click on matlab, then Utility, and there you will find routine set_depth.m which calls stretching.m and these will compute the z-coordinates. You choose whether you get the rho points or w-points coordinates on output. You will find these routines probably call several other functions from this repository, so you might want to just grab everything.

These routines are kept consistent with the Vtransform and Vstretching options available in the myroms code. (It is possible that if you have a very old version of the RNT tools that not all vertical stretching options are supported).

[aearamos]

Hi Pooran,

I've used John Wilkin's tools to compute depth layers on ROMS' output. It has the roms_zslice.m function which can be very useful to you. I've just made a loop with that function and created a new file with the depths I wanted.

Hope it helps.

Arthur

[pooran]

dear friends
many many thanks
i will try them
cheers
pooran

[pooran]

Dear
As you know by using set_depth.m, number of z levels will be same as number of sigma layers. In the other hand each sigma layer has thickness in vertical direction. So please tell me that computed z values for each layer are corresponded to maximum or mean depth in each sigma level.
Cheers
pooran

[pooran]

Dear
i have used set_depth.m and then checked z values in bottom layer with amount of h which are in grid file. There were different. For example for one grid point which its maximum depth is grid file is 227.1217 m , computed z from surface to bottom are, -0.0927,-.2868,....,-10.1409,-44.2666,....,-166.8571,-205.3641, respectively. it means this routine gives -205.3641 m as maximum depth of this cell and also surface z is not equal to zero. also for cells with hmin values, amount of caluated z are not equal to fouth column of abow table With regard to:

Perhaps you can use the ROMS output that looks like this to compare to:

Code: Select all

level   S-coord     Cs-curve   Z   at hmin       at hc    half way     at hmax
32   0.0000000   0.0000000        0.000       0.000       0.000       0.000
    31  -0.0312500  -0.0000256       -0.101      -0.094      -0.198      -0.211
    30  -0.0625000  -0.0001040       -0.202      -0.188      -0.421      -0.474
    29  -0.0937500  -0.0002400       -0.304      -0.282      -0.672      -0.792
    28  -0.1250000  -0.0004423       -0.406      -0.376      -0.956      -1.175
    27  -0.1562500  -0.0007236       -0.508      -0.471      -1.278      -1.635
    26  -0.1875000  -0.0011014       -0.610      -0.566      -1.647      -2.188
:
:
     1  -0.9687500  -0.8117684       -6.190      -5.342    -400.420    -794.233
     0  -1.0000000  -1.0000000       -7.000      -6.000    -492.123    -977.246

if compartion with ROMS out put is in following way:
0*277.1217=0
-0.0312500* 277.1217=−8.660053125
-0.0625000* 277.1217=−17.32010625
:
:
-0.9375000* 277.1217=−259.80159375
-1* 277.1217=- 277.1217

please tell me why those values which calculated by set_depth, calculated from ROMS out put are different with h values is grid file? What is my fault? it makes me confused?
i become so grateful to receive your command

in my grid file
Vtransform=2;
Vstretching=4;
theta_s = 8.0d0
theta_b = 0.4d0
hc=6;
N=32;

[wilkin]

The issue is explained in this depiction of the vertical coordinate definitions:
https://www.myroms.org/wiki/index.php/F ... l_grid.png
and in the background on the vertical discretization at
https://www.myroms.org/wiki/index.php/N ... retization

The "rho" points depths are the vertical center of the cells (or layers). The interfaces are at the "w" points. Hence there are N rho-points and N+1 w-points. The bottom of the bottom-most layer is z=-h, and therefore the z-center of the bottom-most layer is above this. Likewise, the top of the top-most layer is z=0 (actually z=zeta) so the z-center is below this.