Ozone Ancil grid spacing error using Xancil

Hi,

I’m trying to run simulations with prescribed ozone (stash_req = 60). I have taken a 12 month climatology of UKCA O3 (stash 34001) with shape (12,85,144,192) and used xancil to create an ozone ancil (using the ozone subsection in the Atmosphere Ancillary Files list). UM model version number is set to 11.2 (suites I am using range from vn 11-12), atmosphere grid is ENDgame, I specific 85 model levels and 85 ozone levels and for the vertical file namelist file I use one which has successfully generated dump files for UKCA species.

When I try to run with this ancil, initially the model fails with the below message - setting l_ignore_ancil_grid_check=.true. turns the error into a warning and allows the model to run.

rose-app.conf entry

  • [namelist:items(d7ac3352)]

    ancilfilename=‘/work/n02/n02/jweber/INHALE_ancils/dr717_34001_clim.anc’

    domain=1

    interval=30

    l_ignore_ancil_grid_check=.true.

    !!netcdf_varname=‘unset’

    period=3

    source=4

    stash_req=60

    update_anc=.true.

    !!user_prog_ancil_stash_req=

    !!user_prog_rconst=1.0e+20

Error / Warning

  • ? Warning code: -30000

    ? Warning from routine: ANCIL_CHECK_HORIZONTAL_GRID

    ? Warning message: Mismatch between model and ancil field x grid spacing

    ? Expected x grid spacing : 1.8750000000

    ? Ancil x grid spacing : -1073741824.0000000000

    ? Ancil file : /work/n02/n02/jweber/INHALE_ancils/o3_ancils/dr717_34001_clim.anc

    ? Lookup num : 1

    ? Stashcode : 60

    ? Warning from processor: 0

    ? Warning number: 53

Even though the error can be converted to a warning, allowing the model to run, I think I need to fix the issue because, when I compared the dump file for 00:00 1st Jan after a one year run to the mean of Jan and Dec of original netcdf used to create the ancil, there were several clear spatial differences.

It is suspicious that the x grid spacing : -1073741824.0000000000 appears to be one of the default fill values which makes me think it could be issue of a mask somewhere in the ancil generation.

Would you have any advice?

Cheers

James

Hi James,

Is the original ancillary a zonal mean, or is your new ancillary a zonal mean? It might be that it’s only expecting a single X point and so the grid spacing is set to missing data. There is a namelist entry to switch to using a 3D ancillary (zon_av_ozone I think).

Best wishes,

Luke

Thanks, Luke. When I use xconv on my ancil, it has time/altitude/latitude/longitude axes. In the suite (u-dr980) I have zon_av_ozone=.false. and i_ozone_int=1 (3D) so I think everything is set for 3D fields.

Best wishes,

James

Just an update on this for anyone in a similar situation. When I did a 15 day run (360-day calendar), the dump at 00:00 16th Jan matched the ancil monthly mean perfectly.

James

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.