[MITgcm-devel] seaice adjoint and EVP

[Dimitris Menemenlis]

Jinlun, thank you for suggestion.  We've tried time steps down to 1 s.  So it 
must be something else.

On a separate topic, your 2007 AOMIP paper re background vertical diffusivity, 
what ocean advection scheme were you using.  An and I find that reducing 
bacground vertical diffusivity in Arctic does not change much, we suspect 
because background coefficient is swamped by spurious numerical diffusivity from 
the advection scheme.

D.