wiki:ticket/237/Progress_4Nov2015

PeterS, Jin, Susan and Tan discussed progress on this today, and follow up discussions with Lawrie

Priorities

  1. Single run of UK system
    • as close as possible to UK settings, Met Office output.
    • will need local BUFR tables etc. and other local tailoring
  2. Local run
    • as for step1 but with BoM output and other modifications as required
  3. Cycling
  4. Archiving
    • no point archiving until after cycling
    • Tan has suggested directory structure for OpENDAP at NCI
      • netCDF files under OpENDAP
      • start-dump,LBCs, BUFR files archived separately
      • can modify MARS to access these
      • ISS Build Team generalized data access layers should also make use of this
  5. Diagnostics
    • rely on archive

CreateODB

Satellites

  • Tan provided fix to Xiao so that CreateODB correctly handles BUFR Section2
    • Passed on to David Davies, in the trunk
    • Originally BUFR decoder did not read the header, but delved straight in
    • Also included fix to alloe for CrIS data to be read

Insitu Obs

  • UK BUFR tables are out of date and inconsistent with ECMWF
  • We should use ECMWF BUFR_TABLE_B
    • this can b managed by an environment variable
    • fix for this is not in the trunk, as still needs testing
    • Will need to use our own Element Files

To Do

  • AMVs still a couple of days away (Tan)
    • probably a problem with the Quality Indicators.
  • High rejection rate from sondes (Tan, maybe Tomasz?)
  • MTSATclr to be checked (Jin)
  • WindSAT to be checked (whoever of Tan, Jin, Susan is available first)
  • Check Adam Maycock's ticket61 for his progress on this.
  • Will need something to automatically check for changes in UK Element Files, so that we can update our Element Files if needed
    • THIS NEEDS TO BE DONE AS PART OF STEP 1 (Single run of UK version)

MOGREPS

  • Susan helping David extract necessary data from MASS

VAR

  • Susan has it crashing. Looks like a build problem (compiler and/or MPI libraries)

UM

  • Susan has the IO-Server crashing - also looks like MPI library or build problem

Output data

  • There are fields the UK put out that we do not use : not a problem
  • There are fields we put out (variable and/or frequency) that the Met Office do not write or do not check - these are a problem
  • So will always need our own STASH and output settings
  • Modifying our output files means checking all downstream systems
  • So need scripts to automatically check for UK changes to STASH or output file settings, and modify ours as necessary
    • THIS NEEDS TO BE DONE AS PART OF STEP 2 (Single run of Local version)
Last modified 5 years ago Last modified on Nov 5, 2015 9:34:47 AM