Opened 2 years ago

Closed 2 years ago

#2679 closed help (fixed)

Ancil file timings

Reported by: ChrisWells Owned by: um_support
Component: UKCA Keywords:
Cc: Platform: Monsoon2
UM Version: 10.9



I've been trying to run the UM with repeating emissions (1 year repeating), but I can't get it to work.

I get the error

???!!!???!!!???!!!???!!!???!!!       ERROR        ???!!!???!!!???!!!???!!!???!!!
?  Error code: 13
?  Error from routine: GET_EMFILE_REC
?  Error message: Invalid time record 2, End-of-file reached ?
?  Error from processor: 0
?  Error number: 39

which I think is related to my time metadata in my ancil files, but I have a broader question - how does the model deal with the time data of input files, especially for ukca ancils, but also for start dumps? e.g. the start dump referred to in the files is from 1988 but works with the model basis time set to 2014.

Many thanks,

Change History (5)

comment:1 Changed 2 years ago by ChrisWells


So I've been digging, and from looking at trac (#2367) and doing some tests, I see that it doesn't like that the nc files are 12 months (in my suite u-bd260; atmos only with UKCA) - hence the error code 13. What I don't understand is that some of the files I use are not within the time frame I am running in, but the model doesn't seem to mind. Specifically, e.g. for DMS I have the file /projects/um1/ancil/atmos/n96e/ukca_emiss/biogenic/v1/

which is only for 1992, but the model doesn't mind. Is this because I have another DMS file in the time period of the run, and if so, does this mean the model may not be reading in the 1992 data?

I also don't see how the model deals with start dumps - the atmos start dump that I think my run is using is from 1988, and I don't see how it converts that to use at the start of the run - does it just transplant the data to the start time?

In general my question is: is it just for nc files that the date matters? And if so, does my run use e.g. the 1992 DMS file, and in general any files which don't have the date of the simulation? Does the start dump work as expected despite not being labelled the right date? and does the same occur for sea ice/SST?

Very grateful for any help on this, or pointing to relevant documentation,

Many thanks,

comment:2 Changed 2 years ago by willie

  • Component changed from UM Model to UKCA
  • Platform set to Monsoon2
  • UM Version set to 10.9

comment:3 Changed 2 years ago by ChrisWells

Sorry for another update.

In /home/d00/chwel/cylc-run/u-bd166/log/job/20110601T0000Z/atmos_main/01/job.out , I can see that the only ocean ancil files seem to be 1 each for sst and seaice. Both are daily files from year 2000, climatology 1996-2005 - am I right to think that this will just repeat for each year of the model (being as it's atmos only)?


comment:4 Changed 2 years ago by ChrisWells

Found the answer to this - the attribute "update_type" in the nc files determines if they're repeated annually or not.


comment:5 Changed 2 years ago by ChrisWells

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