error when using new sst+ice ancils in HadGEM2-A run

[jonny]

Hi,
I have taken the standard HadGEM2 N96 AMIP run modified this to use cyclical ancils for a time-slice run. This works fine with one set of SST and sea ice ancil files from Buwen Dong (xmodb). But when I try and use my own it does not (xmodd). The ancil files are here:

/work/n02/n02/jonny/ancil/amip_thin/ice_ctrl_AMIPII_nomask_eqarea
/work/n02/n02/jonny/ancil/amip_thin/sst_ctrl_nomask_eqarea
and .leave file is here:
/home/n02/n02/jonny/um/umui_out/xmodd000.xmodd.d17012.t120326.leave
and attached

[jonny]

.leave file from faild run

[jonny]

Am I correct in thinking that the model is failing because of failure of convergence in the numerical solution, rather than due to an issue with the ancil file structure itself. i.e. it is reading the file OK, but falling over potentially because of it's contents?

Thanks,
Jonny

[grenville]

Its hard to tell - can you run again but switch in Extra diagnostic messages (Sub-model indepen..→ output choices)

Grenville

[jonny]

Replying to grenville:

Its hard to tell - can you run again but switch in Extra diagnostic messages (Sub-model indepen..→ output choices)

Grenville

I've rerun it with that switch on.

Cheers,
jonny

[jonny]

The xancil job file can be found on met-cluster at /export/cloud/amh/users/jonny/ancil/amip_thin/ctrl_nomask.job

[grenville]

Jonny

Was this resolved or progressed off line?

Grenville

[grenville]

Hi Jeff,
some combination of running the standard AMIP job to get a new dump file and making those modifications to the ancil files has worked and I have run half a year or so, with my ancil files.

Thanks for your help
Jonny