search

ACCESS-NRI / dev_coupling

0 stars 0 forks source link

Use a Jan atmosphere restart for a better match with forcing #16

Closed MartinDix closed 1 month ago

MartinDix commented 4 months ago

Current suite, e.g. u-db245, starts from Jan 1 but is using the Sept 1 UM restart file /g/data/access/TIDS/OCEAN/hadgem3/initial/atmos/N96L85/ab357a.da19810901_00. This is because the suite was derived from a Met Office AMIP suite which conventionally starts in September.

MartinDix commented 4 months ago

Used Jan restart from the GAL9 AMIP6 run anitial=u-cv124 /g/data/p66/mrd599/cv124a.da19820101_00. Also requires change to use model_basis_time=1981,01,01,00,00,00

Model ran as before, ok with calc_tsfc = .true., crashed quickly with calc_tsfc = .false.

kieranricardo commented 4 months ago

@MartinDix weirdly this is causing things to crash more quickly. Maybe the Jan restart -> more ice -> CICE fails earlier?

kieranricardo commented 1 month ago

Working now!