REGRID - input gridded data does not contain model grid:

[turuncu]

Hi,

I try to run ROMS for Mediterranean region. I have already setup grid and input files but when i run the model, it gives,

REGRID - input gridded data does not contain model grid:

Gridded: LonMin = -12.83264334 LonMax = 44.14562002
LatMin = 27.69616640 LatMax = 49.65302851
Model: LonMin = -12.91627787 LonMax = 44.22968287
LatMin = 27.68158081 LatMax = 49.65305395

GET_2DFLD - error while reading variable: Uwind at TIME index = 124

and exit. I check the input grid file and forcing file and the both of them has is same grid coordinate (RHO, U, V). In grid file, spherical option is also set to 'T' but model still gives same error. In fact that, i don't want to interpolate ant data and i think that REGRID is used for interpolating coarse data-set into ROMS grid. How can i solve it? Is there any CPP option?

PS: i set the x_rho, y_rho, x_u, y_u, x_v, y_v to lat lon values. Because of the 'T' definition of spherical variable. Otherwise, model try to use projected coordinates.

Thanks,

[jcwarner]

check again.
Because the error is saying that the gridded wind data does not cover the entire ocean grid. For example:
Gridded: LonMin = -12.83264334
Model: LonMin = -12.91627787
so you need more wind data to the left.

[kate]

Also, ROMS doesn't use a cpp option for this, but rather the grid dimensions. Do the values of xi_rho and eta_rho match up between the grid and forcing files?

[turuncu]

I check the grid dimensions and they are same. May be i am little bit confused, so please tolerate me but i think that the error is expected when i use the lat-lon rather than x-y because lat-lon is in Lambert Conformal projection. See the attached plots that represent the grid points location (lower/left corner). The color is same with horizontal grid illustration on

https://www.myroms.org/wiki/index.php/N ... _Technique

it simply represents the rho, u and v points. So, the minimum and maximum values return different for rho and u or v points. Normally, x_rho(1,1) == x_v(1,1) and y_rho(1,1) == y_u(1,1) but in lat-lon case this rule is not provided. I also change spherical = 'F' in grid file to use x-y values but the model gives error like this,

Gridded: LonMin = 484601.88086907 LonMax = 4796398.11912724
LatMin = 185921.38364709 LatMax = 2373078.61635291
Model: LonMin = 0.0 LonMax = 0.0
LatMin = 0.0 LatMax = 0.0

the model x-y data is not read but i don't know why.

[kate]

Again, ROMS is checking to see if the dimensions match - if they match, it won't do any regridding. ROMS is expecting the winds to be at rho points. I have an old wind file on the ROMS grid and here is the ncdump -h on it:

Code: Select all

netcdf Uwind_48 {
dimensions:
        xi_rho = 226 ;
        eta_rho = 642 ;
        wind_time = UNLIMITED ; // (941 currently)
variables:
        double wind_time(wind_time) ;
                wind_time:long_name = "surface wind time" ;
                wind_time:units = "days since 1899-12-31 00:00:00 GMT" ;
                wind_time:field = "wind_time, scalar, series" ;
        short Uwind(wind_time, eta_rho, xi_rho) ;
                Uwind:long_name = "10m Uwind component" ;
                Uwind:units = "m/s" ;
                Uwind:field = "Uwind, scalar, series" ;
                Uwind:time = "wind_time" ;
                Uwind:scale_factor = 0.00228988987869727 ;

// global attributes:
                :type = "NEP4 FORCING file" ;
                :history = "NorthEast Pacific" ;
}

This is dimensioned exactly the same as the Tair file for a grid with Lm=224 and Mm=640.

[turuncu]

Okay, now it makes sense because, i interpolated wind components to u and v points. I will change my code and re try. I hop it will work in this time. Anyway, i think that this point is not clear in documentation but i will check it again. In ROMS forcing page sustr and svstr are in u and v points (simple ncdump output) so i think that wind must be in u and v points. May be table 1 can be modified and new column can be added that contains the actual grid points for each variable. In this case, are all variables in forcing file in rho points? BTW, I am using BULK_FLUX.

Thanks for your help,