Skip to content

December2015

Jump to bottom
rml599gh edited this page Dec 13, 2015 · 48 revisions

CABLE committee meeting, 10th December 2015, 2.00pm

  1. End of year ticket review
  1. Trunk updates:

a) Updates in progress

i) Medlyn stomatal conductance, #56. Minor fixes based on review by BP. JS currently has problem with this version crashing at start of ACCESS run. Will commit to trunk when problem identified.

ii) #62 (radiation/albedo bug fix, ASAP), #92 (if BP confirms), #66 (initialisation bug), #89 These relatively simple changes were checked as part of the ticket review and can be bundled together to go into the trunk after #56

ii) SLI and POP, #18, #61, #93, #94. Progress on remaining issues from last minutes? Query #90/#91/#73? Confirmed that #91 is a duplicate of #90 and can be closed. RL to action. #90 should be already dealt with in the SLI/POP branch and so will be part of the bundle to be included. Need to start testing impact of any of the minor changes (#90, #93, #73) on ACCESS cases (i.e. if switched on). Need to alert BP to #94 work that is I/O related to avoid duplication of effort. In particular, there is useful capability to run a subset of the globe using a global input file.

iii) #95 (and #44) Note related work under Haverd2013 switch (#93). Need to work towards systematic testing to determine best choice of switches for different applications (see below re Ned's possible contribution.

b) Update still to come #70 (BP to check backwards-compatibility)

  1. CABLE for CMIP6, ACCESS-CM2 Video conference held 8/12/15. Discussed timelines (need code coupled reliably by March 2016). Work progressing at UNSW to test bundle of hydrology improvements in ACCESS C20C+ runs.

  2. Land use change in CABLE, LandUse Workshop held 3/12/15. Timelines for implementation into ACCESS for CMIP6 may be challenging. VH proposed initiating some technical discussions early in the new year, and could visit Aspendale if required.

  3. CABLE-JULES coupling

a) Progress with coupling into UM8.5 for ACCESS-CM2. Identifying issues with the runs, but not yet many causes. Post meeting update: change in dynamics to ENDGAME means part of atmospheric code is iterated which was not before. Need to work out what the implications are for how/when CABLE is called.

b) CABLE in JULES4.2/UM10.1, RL meeting with Martin B RL visited Met Office in November. No replacement for JULES technical position yet. Some progress with removing land variables from UM dump but proving more challenging than initially expected.

  1. Benchmarking/test suite

a) CABLE technical test suite – other test users?, document on wiki? CASA-CNP test? No progress. Having difficulty getting BP set up to run ACCESS.

b) PALS Some progress. Might try and build CABLE science testing suite into PALS.

c) Promotion/use of other bench-marking tools e.g. iLambs, ESMValTool Discussion of ESMValTool, given CRESCENDO project (EU) continued development of this tool. PLUME experiment (including CABLE submission) will be evaluated using iLambs.

  1. Technical support

a) Possible role for Ned - follow-up? Three areas of interest to Ned. (1) Science testing suite (encouraged by GA), (2) Transition to Git, (3) standardise code for multiple compilers etc. Committee strongly supports (1) and (3) but has concerns about the community impact of (2) and do not see this as a high priority. Science testing suite would likely be script based, taking a particular revision of the code and running through optional number of increasing complex test cases. The aim is to have a tool which can be used to easily test the range of options available in CABLE for compatibility etc. Next steps are to confirm possible funding with AP, Ned to revise proposal document and speak to us at next meeting.

  1. Meetings

a) LandMIP meeting (October, Zurich), GA to report GA found meeting useful. Discussion of different global forcing datasets. Three sets considered essentially equally good, WFDEI (WATCH-Forcing-Data-ERA-Interim), GSWP3, Princeton but noted that could give quite different results when driving a land model. CRUNCEP also worth considering for investigation since it has been used widely as well (e.g. for MsTMIP). Potential use in LS3MIP. Would be good to set up CABLE to easily switch between these datasets. GA has a student potentially working towards this (Sanaa Hobeichi), and also might be something Ned might engage with regarding CABLE science testing. Link with BP and Lars on potential I/O clean-up.

b) OzeWex (December, Gold Coast), report Small and short session. Not much directly relevant to CABLE.

c) AMOS (special session, land/CABLE, February, Melbourne) Those requesting oral presentations were all able to be accommodated.

  1. Next meeting date: January or February? February 4th, 2pm

  2. Any other business

JK has added extra information to the Beginner's Guide

Clone this wiki locally