I copied u-dq607 and created u-ds461. I thought it ran fine although it was several weeks ago. These are UM vn13.8 global suites.
Now, based on this I’ve been trying to create a suite for future simulations. In u-ds482, I specified CMIP6 input data for 2014-2101 instead of 1850-2014. Because I couldn’t find any of these data I use files stored in my directory. Some other files I still can’t find.
So I don’t expect that the suite will run successfully but I think this error may not be related.
The run failed with this error;
? Error from routine: RCF_CALC_TILE_MAP ? Error message: Consider l_rcf_init_flexi = T or set tile_map_ids: Cannot map tiled field 229, tile id 1
So I set l_rcf_init_flexi = .true. and ran again. It failed with this message;
? Error from routine: RCF_RESET_DATA_SOURCE ? Error message: Multilayer snow recon (rcf_ml_snowpack) is not compatible with l_rcf_init_flexi. Please use tile_map_ids instead.
What can I do? If this is related to the fact that some of the input data are still for 1850-2014, all I need to do is find those data (atmospheric oxidants data). But it doesn’t appear to be the case.
I’m working on upper troposphere and lower stratosphere, and so treatment of snowpack is not extremely important. If there is a way to circumvent this error that would be fine.
This combination might required because of the error message:
Multilayer snow recon (rcf_ml_snowpack) is not compatible with l_rcf_init_flexi.
this means you cannot have both features active otherways it seems that the snow reconfiguration stays on by default.. Hopefully with this it will run, let me know what you get.
Thanks for your help, but it doesn’t seem to work. I got this error;
? Error from routine: CHECK_IOSTAT ? Error message: ? Error reading namelist recon_technical ? IoMsg: The variable name ‘RCF_ML_SNOWPACK’ is unrecognized in namelist input. ? Please check input list against code.
I think I get this error because rcf_ml_snowpack is not a namelist variable. It is actually a module and subroutine. src/utility/qxreconf/rcf_ml_snowpack_mod.F90 contains SUBROUTINE rcf_ml_snowpack.
But I don’t think code change should be necessary for this. I tried running the original suite u-ds461 again and recon finished without an error. The difference between two suites is that u-ds461 is a run for 1988 while u-ds482 is for 2015. So I replaced historic chemistry and aerosol emission data with future ones in app/install_ancil/rose-app.conf. I also changed GHG concentrations in app/um/rose-app.conf.
One of these changes is likely to be somehow causing an error in recon. If you have any idea please suggest.
It will help in debugging if you commit your changes to the suite, but this error may be due to the fact that you are using an UKESM dump to initialise a GAL9 configuration. The land surface between the two models is quite different and hence reconfiguration is getting confused with the surface tiles setup.
It might be better to find a suitable dump for 2015 (or other 2010 years) from an GAL9 run.
Also, re: your earlier query, the central ancillary folder on Monsoon3 is /common/share, so there should be suitable ‘future’ ancils under this folder.
Sorry, I totally forgot about committing. It’s taking me so many figuring out to run a suite on Monsoon3 with cycl 8. Also I got an error when I tried to commit something before, so I may need to fix something before I can commit a suite and a branch.
Anyway, thanks for your advice. So the dump file I was using is incompatible. I’ll ask Anthony if he has any dump for more recent year – well, he doesn’t seem to have one. Do you know where I can find a proper dump file? Or if I start from 1988 dump, how many years do you think I should run to spin up the dump for 2015?
Now I found UKCA input files in /common/share/monsoon_ancils_cmip6/n96e/ssp370/AerosolChemistryEmissions/v1/
I couldn’t find any at all when I searched before for some reason. I looked so hard and I thought I had checked all possible places at that time.
Can you also help me find ozone and oxidant data? I think data like below are used in the suite (specified in app/install_ancil/rose-app.conf ) and they exist, but I can’t find equivalent data for 2015- ;
/projects/metoff/umdir/ancil/atmos/GC5/n96e/orca025/ozone/cmip6/v1/mmro3_monthly_CMIP6v2_clim_3d_1990_2010_ants.anc
and
/projects/metoff/umdir/ancil/atmos/GC5/n96e/oxidants/ukesm1-1_hist1_u-by791/clim_1979_2014/v1/ukca_oxid_clim_*.nc
where * = H2O2, HO2, NO3, O3, OH
All the Standard GA configurations are run for 1988-2008 in any case. I have copied the 200809 dump from u-do527 (parent of u-dq607) to /data/scratch/mohit.dalvi.mon/temp/. You can run a copy of u-dq607 starting at this date upto end-2014 since all ancillaries used should cover the CMIP6 historical period.