Hi all,
does ACCESS-OM2-01 have a dzt (t-cell thickness) grid variable but on the u-grid?
Currently, I am interpolating dzt onto the u-grid using
dzt_interp = dzt.interp(xt_ocean = a.xu_ocean.values, yt_ocean = a.yu_ocean.values,
method="nearest").rename({'xt_ocean': 'xu_ocean', 'yt_ocean': 'yu_ocean'})
… but maybe this is not the best approach?
I don’t think we saved any dzu diagnostics, but if you’re doing new runs you could save it.
How you calculate it offline depends what you’re trying to do. If you just want to do depth averaging then a simple interpolation might be good enough. But if you are trying to close budgets, you will need to recreate what the model does for dzu, which is to calculate the minimum of dzt from the 4 surrounding t-cells.
Ah ha. When I saw @mauricehuguenin’s question I was wondering if it even made sense to have dzt on a the u grid, as that is the boundary between cells. I had a mental model of a blocky ocean that looks a bit like lego blocks of different heights.
I have learnt something new today, thanks @adele157
Thanks for the help @adele157 and @Aidan . I am running a short test case now saving the dzu variable.
If you’re running test cases anyway, couldn’t you just output the end heat flux diagnostic that you want anyway?
Good point, thanks Ryan. With this test run I wanted to estimate how much the difference between the online calculated dzu and my interpolated dzt from above is. The difference is in the order of 0.014% so in my case, interpolating dzt onto the u grid is okay.