Minutes of the QUEST-ESM NCAS FAMOUS meeting of Tuesday, 12th June 2007
Present: Manoj, Simon, Jeff, Robin, Annette, Ros, Jonathan
Annette has made progress with the coastal tiling. After she has manually resolved overlaps with a radiation mod (adb1f406), an automatic procedure can be to used with merge and nmodex to transform the coastal tiling mod for 5.2 into a mod for 4.5 + MOSES 2.2. However, careful checking is necessary. Annette has mostly dealt with the radiation sections and will proceed to the boundary layer. It seems likely this will be completed within a month.
Robin has written a mod for very long absolute time; this writes the year as a signed 9-digit integer (decimal, no letters) in the UM output filenames. He will add a logical switch in a control namelist so that we can include the mod unconditionally in the standard FAMOUS job. It also requires some small executables to be rebuilt to accommodate the longer filenames in history file namelists and printout. Jonathan will test that the modified executables are OK for jobs without the mod; if so, they can replace the standard small executables. Once that is done, Ros could add a UMUI switch to enable the new convention.
Use of long years (back to 1 Myr BP) will be useful for glacial cycles, when calendar years affect the orbital forcing. Robin wrote a mod to calculate the solar coefficients from the present year based on Berger (1978) for orbital parameters. This is more convenient than the one which Julia Tindall of Bristol helpfully supplied, since that one requires an external data file to be read. Manoj pointed out that a similar facility was implemented in the UM after 4.5 by William Ingram. Robin finds that his mod and William’s are similar, so he will adapt the 6.1 code for 4.5.
Robin’s treatment of tracer conservation, with volume-average correction, appears to work for the DIC and Alk as well. He is awaiting an answer from Chris Jones or Ian Totterdell about a warning concerning these diagnostics in the stashmaster file.
Robin has retuned sea ice albedo and new ice thickness. Along with the deepened bathymetry, this practically eliminates the cold N bias. The ozone-shifting mod continues to be problematic.
The new UM server machine has arrived. Ros has installed the UMUI and its database. The UMUI has the option to make experiments private, and to grant access to named users. She expects that some Reading users will be invited to try it next week. In that case, it ought to be available to users outside Reading, who will be given individual usernames, some time next month. Possibly the UMUI databases from elsewhere could be imported into the new one, if the initial letters of the experiments were changed from x. The UM helpdesk is also available on the server, and can be used as soon as users have accounts.
Jeff has made some progress with xancil but has had little time. He thinks that two months from now ought to be sufficient. He would like to have finished the work before September.
Apart from ancillary files, we will have delivered all of QUEST-ESM D2 in some form once the UM server is available to external users.
Simon is working on a hand-edit to enable submission (with umsubmit) of jobs on the QUEST cluster. He has run a 1-day HadGAM1 at N48 and will do a longer timing test; this will be of interest for QESM. Over the next month he will try various timing tests, including N96 HadGEM1. Gethin Williams will copy the UM installation from ormen to quest once Simon is ready; that will be in the next week. Once that is done, Annette will run a test FAMOUS on quest.
There is a FAMOUS user meeting in Bristol on Friday, organised by Andy Ridgwell. Robin is going. Jonathan might be able to.