Noise in Jnr W

W at hybrid height of 5,600m

Old run at 3 day dump	Old run at 3.25 days
New run at 3 day dump	New run at 0.25 days

The problem was the orography was being passed too late!

The order of calls was

• -> INITIAL_4A
  • -> SETCONA_CTL_4A (where orography metrics are calculated)
  • -> OASIS_INITIALISE_2 -> OASIS3_GET_OROG_JNR

Answer is to move call to OASIS_INITIALISE_2 above SETCONA_CTL_4A (I really should have spotted this earlier - it was done properly at UM8.6)

All the arguments about vertical levels in Jnr not matching are still true

We should still look at this, but on back-burner for moment.