map river nutrients to blom mesh using rof->ocn mapping files

[mvertens]

This PR implements mapping river nutrients on a 0.5 x 0.5 "runoff grid" and mapping the nutrients to the blom grid using the offline mapping files for mapping rof->ocn.

The river nutrient file is here
/cluster/shared/noresm/inputdata/ocn/blom/bndcon/river_nutrients_GNEWS2000c00_r05_20250220.nc

This has to be tested with #534. @JorgSchwinger - can the data for this new riverine input also be generated on the 0.5x0.5 grid - so that I can test with the updated dataset?

[JorgSchwinger]

This has to be tested with #534. @JorgSchwinger - can the data for this new riverine input also be generated on the 0.5x0.5 grid - so that I can test with the updated dataset?

I have copied a file for the R2OMIP river input data (same format, same grid as GNEWs but contains one additional variable) to

/cluster/shared/noresm/inputdata/ocn/blom/bndcon/river_nutrients_R2OMIPpi_r05_20250527.nc

Please let me know if there are any problems with the file.

[jmaerz]

Hi @mvertens , why do you repeat the fldlist_input here (see l. 54)?

[mvertens]

Good catch. I fixed this.

[jmaerz]

This currently seems to only support the 'old' river input file - for the river2omip the slDOC seems to be missing.

[mvertens]

That is correct. @JorgSchwinger just provided a new input dataset that I can use and I'll add that field.

[jmaerz]

== 'DET' ?!

[mvertens]

This is now fixed as well. Thanks.

[TomasTorsvik]

@mvertens - Thanks, looks good to me!

[mvertens]

This has to be tested with #534. @JorgSchwinger - can the data for this new riverine input also be generated on the 0.5x0.5 grid - so that I can test with the updated dataset?

I have copied a file for the R2OMIP river input data (same format, same grid as GNEWs but contains one additional variable) to

/cluster/shared/noresm/inputdata/ocn/blom/bndcon/river_nutrients_R2OMIPpi_r05_20250527.nc

Please let me know if there are any problems with the file.

@JorgSchwinger - thanks for providing this. I don't see an entry for DET in this file. Yet in mo_read_rivin.F90 the following is called - and DET is always read in.

    call ncfopn(trim(rivinfile),'r',' ',1,iotype)
    call ncread('DIN',riv_DIN2d,dummymask,0,0.)
    call ncread('DIP',riv_DIP2d,dummymask,0,0.)
    call ncread('DSi',riv_DSI2d,dummymask,0,0.)
    call ncread('DIC',riv_DIC2d,dummymask,0,0.) ! It is actually alkalinity that is observed
    call ncread('Fe' ,riv_DFe2d,dummymask,0,0.)
    call ncread('DOC',riv_idoc2d,dummymask,0,0.)
    if (use_river2omip) then
      call ncread('slDOC',riv_itdoc2d,dummymask,0,0.)
    else
      riv_itdoc2d = 0
    endif
    call ncread('DET',riv_idet2d,dummymask,0,0.)
    call ncfcls

In addition, is BLOM master - what file is used if use_river2omip is .true.? Is this currently just set in user_nl_blom?

[JorgSchwinger]

@mvertens, yes, you are right the DET field is incorrectly named 'POC' in the file

/cluster/shared/noresm/inputdata/ocn/blom/bndcon/river_nutrients_R2OMIPpi_r05_20250527.nc

@TimotheeBrgs can you please fix this such that we have identical names for the GNEWS and the R2OMIP cases?

(For the interpolated files I fixed this in the interpolation-script, but it would be better to have this correct from the beginning).

And yes, if you use R2OMIP, you currently you need to specify the input file manually in the user namelist - something that has to be fixed once this option is tested and finalized.

[TimotheeBrgs]

Hi @mvertens I have now fixed the wrong naming of the DET variable. I replaced the former file with the one below. Hoping that everything else is fine.

/cluster/shared/noresm/inputdata/ocn/blom/bndcon/river_nutrients_R2OMIPpi_r05_20250604.nc

[mvertens]

@TimotheeBrgs - Thanks!