UNRECOVERABLE library error

1

Hi,

I’m trying to run coupled suite u-cm339, a copy of the vn12.0 historical release job, on Monsoon but the suite fails on the coupled task with the error:

Encountered during a sequential formatted WRITE to
lib-4171 : UNRECOVERABLE library error
An output list item is incompatible with its data edit-descriptor.

I have made some changes to u-cm339 so that it can run from 2015 onwards (as an SSP variant) and use the CRI-Strat 2 mechanism and corresponding wider set of emissions. The suite runs fine (with CS2 and extra emissions) when I use the vn12.0 trunk. However, when I use my branch vn12.0_CS2_SOA_improvements_w_ST_DMS_v3, it stops on the 3rd timestep, which leads me to think it is a problem with the chemistry. My branch works fine with AMIP suites (e.g. u-cl374).

I have found one previous instance of this error (http://cms.ncas.ac.uk/ticket/1637) with the solution relating the time matching of the CRUN re-initialised pp file. However as my suite runs fine with the trunk, I’m not sure if this is relevant in this case. I’m also using a 2015 dump file to initialise the atmosphere.

When I had this UNRECOVERABLE error some time ago I fixed it by removing some incorrectly written WRITE statements in the branch I was using at time. I have removed/commented out all of my WRITE statements in my current branch and checked that the suite is using the correct version of the committed branch yet the problem remains.

Is there anything else which could cause this type of error?

Many thanks for your help,

James

2

James

The problem is in ukca_aero_ctl.F90 - the WRITE statement is wrong, but I guess the real problem is
that msecorgi is <= 0

   cmessage='msecorgi <= 0' ! This appears to be causing some problems
    WRITE(umMessage,'(A20,I6)') cmessage,' msecorgi = ',msecorgi
    CALL umPrint(umMessage,src='ukca_aero_ctl')

Grenville

3

Hi Grenville,

Thank you for spotting this. I realise now that my branch has to be run with the non-standard option of i_mode_setup = 4 since I have added an extra Sec_Org tracer (Sec_Org_i) which is only considered by the aerosol scheme when this option is selected. My suite now runs.

James