Opened 3 years ago

Closed 3 years ago

#1999 closed help (fixed)

Xancil for 3D user ancillary file

Reported by: earhg Owned by: um_support
Component: UM Model Keywords:
Cc: Platform: MONSooN
UM Version: 8.4


I have a netcdf file of ion production rates from radon which I have regridded to N96L85 and I would like to load them into UKCA vn8.4 (I don't need a radon tracer, just monthly mean values). I thought (maybe incorrectly) the easiest way to do this would be to follow the UKCA Tutorial 5 to add this to the existing 3D emissions which are read in from an ancillary. When I try to make a new multi-level user ancillary (/projects/clarify/hamgo/ancils/ancilcreate.job on Monsoon) I find the existing 3D fire and aircraft NOx emissions from GFED3_2002-11_AC.N96L85nd have 84 model levels ("molality") and 1 level ("mass flux") instead of 85 model levels. Do I need the 'mass flux' part of this in my new ancillary, and how should I get my radon file to be compatible with these? Or should I do this a different way? Many thanks for any help - Hamish

Change History (5)

comment:1 Changed 3 years ago by earhg

(to clarify this, since in Xancil the grid configuration is set once (I use /projects/um1/vn8.4/ctldata/vert/L85_20m_85km_15_6km_qs_o1) the grid for my ion production rates has to be the same as the other ancillaries, otherwise I can't load it. But I thought these would have 85 levels, not 84…

comment:2 Changed 3 years ago by earhg

hi again,
I see by comparison of the level_height coordinate with other datasets that the bottom (surface) model level is missing from these 3D emissions files. However, it is important to keep this level in my ion production rate file - its where most of the radon comes from. To keep the format of the ancillaries the same, I guess I can remove the TOP (85th) model level from my ion production rate file (which is all zeroes anyway), read the file into UKCA with 84 levels as if they were the 84 levels from the second level up, and then move all the ion production rates down in the UKCA code afterwards. I will close the ticket if this ends up working….sorry for making noise.

comment:3 Changed 3 years ago by luke

Hi Hamish,

I think that this is either an issue with the Met Office produced ancillary file, which the UM can read fine, but Xconv can't, or an issue with Xconv in that there is something in the file headers that messes up the reading step. (or possibly, its a problem with both. I've only seen this issue on files produced by the Met Office though, not by ones produced using Xancil).

Can you try using Iris (or cf-python, if it can read the ancillary file format) to read the ancillary file and then write out the netCDF file containing the data? This should work I think.

If it doesn't, I have an old fortran script which you can use to combine these emissions back to one field, which would also work. I think that using python is probably best to try first though.


comment:4 Changed 3 years ago by earhg

hi Luke,
Yes, I think xconv screwed up the conversion of the ancillary to netcdf format, and this is why there were 84 levels when the UM needs 85. At first I thought 84 was right, but then when my job didn't run I added a slice of zeroes to the 84 levels in iris to make it back up to 85 levels, merged in my radon, made new netcdfs and processed them with xancil and now my job is running (and will give roughly correct output so long as xancil hasn't screwed up the conversion). I didn't think iris would read the ancillary directly, but I will try doing that and see if I get the surface level this time. Since my job is working, I will close the ticket. Many thanks for your advice! Hamish

comment:5 Changed 3 years ago by earhg

  • Resolution set to fixed
  • Status changed from new to closed
Note: See TracTickets for help on using tickets.