[MITgcm-devel] longtimestep test

[Jean-Michel Campin]

Hi Oliver,

thinking of what would be a good test for the longtimestep pkg:
- to reproduce fairly well the KPP mixed layer from the coupled
  model requires some high frequency forcing, so a lot of 
  files. And would need also to run for few long time-step 
  which means many more dynamical time-step. 
  With a global set-up like cs32 and with a decent vertical resolution,
  this could take some time to run (with all debug options).
- alternatively, we could modify a small (toy-like )set-up by adding 
  ptracers + longtimestep (either change it or adding a second test
  with an input.{something} dir). For example, natl_box (forward)
  is already using KPP, and the forcing files are already there.
  and rbcs could be added at the same time with ptracers.
We can talk on Monday about this.

Cheers,
Jean-Michel