Re: [ferret_users] unknown model resolution in the vertical

[Martin Schmidt <martin.schmidt@xxxxxxxxxxxxxxxxx>]

Hi,
This result should be correct!
Your variable depends on "z" not on "depthu".
variables:
    long time_counter(t) ;
    float vozocrtx(t, z, y, x) ;

The same is valid for your output variable, the horizontal coordinates are changed as it should be, otherwise it depends again on "t" and "t".

What is missing in your new file is the information on the vertical levels.
    double depthu(z) ;

You need to save this variable in your new file and can work with the vertical coordinate like in the original file. Hopefully ...

Best,
Martin
On 03/13/2017 04:09 PM, Arijeet Dutta wrote:

Hi

original file i.e. in native grid dumps

netcdf vozocrtx_fa9p_1m_2001_grid_U {
dimensions:
    t = UNLIMITED ; // (12 currently)
    z = 42 ;
    y = 292 ;
    x = 362 ;

variables:
    long time_counter(t) ;
    float vozocrtx(t, z, y, x) ;
        vozocrtx:_FillValue = 9.96921e+36f ;
        vozocrtx:coordinates = "time_counter depthu u_lat u_lon" ;
        vozocrtx:long_name = "Zonal Current" ;
        vozocrtx:offline_operation = "monthly mean" ;
        vozocrtx:units = "m/s" ;
    double depthu(z) ;
    float u_lat(y, x) ;
    float u_lon(y, x) ;
    byte umask(z, y, x) ;

After curv_to_rect, new file dumps

dimensions:
    XAX = 91 ;
    YAX = 76 ;
    Z = 42 ;
    T = UNLIMITED ; // (12 currently)

variables:
    double XAX(XAX) ;
        XAX:units = "degrees_east" ;
        XAX:point_spacing = "even" ;
        XAX:axis = "X" ;
        XAX:modulo = 360. ;
        XAX:standard_name = "longitude" ;
    double YAX(YAX) ;
        YAX:units = "degrees_north" ;
        YAX:point_spacing = "even" ;
        YAX:axis = "Y" ;
        YAX:standard_name = "latitude" ;
    double Z(Z) ;
        Z:point_spacing = "even" ;
        Z:axis = "Z" ;
        Z:standard_name = "altitude" ;
    double T(T) ;
        T:axis = "T" ;
        T:standard_name = "time" ;
    double U1(T, Z, YAX, XAX) ;
        U1:missing_value = -9.999999999999999e+33 ;
        U1:_FillValue = -9.999999999999999e+33 ;
        U1:long_name = "CURV_TO_RECT(U[D=1], MAP[D=2])" ;

with z values

 Z = 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20,
    21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38,
    39, 40, 41, 42 ;

greetings

Arijeet

On Mon, Mar 13, 2017 at 8:12 PM, Martin Schmidt <martin.schmidt@xxxxxxxxxxxxxxxxx> wrote:

Difficult to say. Please inspect your data with
ncdump -h

This shous you the header section of your data set, especially the coordinate axes and how the varaibles depend on them.
Best,
Martin

On 03/13/2017 02:28 PM, Arijeet Dutta wrote:

Hi all
I have data (ftp://ftp.icdc.zmaw.de/EASYInit/NEMOVAR/) which are in a tripolar native model grid, I am able to interpolate it to regular latlon grid using curv_to_rect. The problem is it has 42 vertical levels and all I know is 15 of which are in upper 200m. Any idea how I can get depth values from it?

greetings,
Arijeet