Hycom-Ncoda in PYROMS

[hpftcb]

Hi all,
I am trying to generate boundary and initial conditions to ROMS from Hycom-Ncoda Analysis-Reanalysis through PYROMS. I am using as background examples/Arctic-HYCOM. I have installed everything correctly (I think ),so I reckon my problem is not related with Pyroms installation. The thing is I am using as input Hycom-Ncoda files that contain everything in it, instead of the way used in the example, in which different variables are downloaded to different files (The link is broken and I did not find another one, so I downloaded directly from Hycom-Ncoda system). The problem is that my output data is bizarre, it is not being interpolated to the right places, and there are lot of NaN (1.26 ...) in my data, in places they should not be; Therefore, is there anyone that has used Hycom-Ncoda analysis/Reanalysis downloaded manually from the website to create boundary condition using PYROMS. I think that some matrices have to be rotated or transposed to fit the new input data format, but I can not figure what to do, nothing works. Some stripes of NaN appears in the interpolated data, it is really strange.
Besides, has anyone the link to download Hycom data in the format used in examples/../get_hycom_GLBa0.08_grid.py?

My ROMS grid was generated through Seagrid.

Thanks in advance,

Fernando Tulio Camilo Barreto

[ptimko]

Are you trying to read native HYCOM (*.[ab]) output files? If it is a global dataset then the data are on a global tri-polar grid, which is not rectangular. There should be FORTRAN routines available on the HYCOM website for converting native HYCOM (*.[ab]) to netcdf (*.nc) which might help. At least then you could then read the data into python as a netcdf file.

[hpftcb]

Thank you for your reply. I've downloaded hycom-Ncoda data in .nc format for an area bigger than my roms area. Pyroms is reading and interpolating, the problem is that the output data is wrong, it is like the program is not interpolating for the right points. At least until after flooding, everything is all right, the problem is in the horizontal interpolation. The input seems fine, perhaps the problem is with weights, I do not know.
Thank you.