Custom Query (986 matches)

The recent IO updates appear to have broken a tidally-forced application of mine. The tidal period and parameters (elevation amplitude, etc) are not being read from the file, as illustrated by the following output:

    GET_NGFLD   - tidal period
                   (Min =  3.11285267-303 Max =  3.11285267-303)
    GET_2DFLD   - tidal elevation amplitude
                   (Min =  1.00000000E+37 Max = -1.00000000E+37)
    GET_2DFLD   - tidal elevation phase angle
                   (Min =  1.00000000E+37 Max = -1.00000000E+37)

The problem seems to be that in ROMS/Nonlinear/get_idata.F at line 101, NTC(ng) (i.e. the number of tidal consttituents) is set to 0

      IF (iic(ng).eq.0) THEN
        NTC(ng)=0
        CALL get_ngfld (ng, iNLM, idTper, ncFRCid(idTper,ng),           &
     &                  nFfiles(ng), FRCname(1,ng), update(1),          &
     &                  1, MTC, 1, 1, 1, NTC(ng), 1,                    &
     &                  TIDES(ng) % Tperiod(1))
#  if defined AVERAGES_DETIDE && (defined SSH_TIDES || defined UV_TIDES)
        TIDEname(ng)=TRIM(ncfile)
#  endif
      END IF

and it never acquires any other value. (In my application it should be 1.) Therefore all the subsequent read requests return 0 data values.

As an ad hoc fix, setting NTC(ng) to 1 in get_idata.F causes sensible tidal forcing data to be read from the file

    GET_NGFLD   - tidal period
                   (Min =  4.47141632E+04 Max =  4.47141632E+04)
    GET_2DFLD   - tidal elevation amplitude
                   (Min =  9.37254056E-02 Max =  1.37402248E+00)
    GET_2DFLD   - tidal elevation phase angle
                   (Min =  2.47094957E+00 Max =  5.79200999E+00)

I don't know how NTC(ng) should get its value, however.

Finally, I am done with phase III of the parallel I/O implementation in ROMS. Notice that most of the calls to the NetCDF library functions (nf90_*) are made from the ROMS module (mod_netcdf). The exceptions are def_dim.F, def_info.F, def_var.F, and reading (nf_fread*d.F) and writing (nf_fwrite*d.F) routines.

For more information please check the following ​link to the forum.

Corrected a bug in nf_fread2d.F when interpolating input fields to model grid using regrid. This bug was introduced during the parallel I/O implementation. It affects the serial I/O. I missed broadcasting input data in wrk array. I need to test how to regrid in parallel I/O.

Many thanks Greg Gerbi for reporting this problem.