What we've done

MEDUSA

  • now fully coupled
    • Units of CO2 flux needed converting in MEDUSA (the conversion may be wrong)

Bug hunting PFTs

9 PFTs

  • The original coupled job crashed soon after starting
  • Started again from a 5 PFT (the standard) coupled run which worked, u-ae029
  • Gradually added in the 9PFT complexity (u-ae988)
    • Job runs for just over 10 days before problem with convection
    • Suggests the original job had some wrong setting(s) in TRIFFID or something (I haven't investigated)

13 PFTs

  • The original coupled job was crashing
  • John Edwards found that we need
    • tile_map=1,1,1,2,2,3,3,3,4,4,4,5,5,6,7,8,9
    • l_rcf_init_flexi = .false. (an obsolete flag)
  • Using `made-up' ancillary data, e.g. C3 crop canopy height = C3 grass canopy height
  • Optimisation: high -> safe

Creating a prototype UKESM0.4

  • Like most other UM10.5 jobs, we wanted to switch to `split app'
  • GC3 at UM10.5 still has EG_BICGSTAB error (u-af082)
  • Instead of GC3, started from 17 tile `split app' job at UM10.4.1 (u-af280)
  • Added in ORCA1 ocean from u-ac926 (u-af344).
  • Upgraded job to UM10.5 (u-af408)
  • Added in MEDUSA (Julien, you probably want to make changes to medusa app) (u-af755)
  • Add in full chemistry BVOC etc (u-af734)
  • Add in Till's updates to ORCA1, except iodef.xml (u-af826)