[Mitgcm-support] Re: grid scale oscillations

[mitgcm-support at dev.mitgcm.org]

Hi,

Unfortunately, none of the suggestions resulted in a significant
improvement. However, I've had some success by changing to a 2nd order
flux limited advection scheme. I am using cosPower=2 as suggested by
Alistair.

What is worrying me about this though is that I am wondering whether the
flux limited advection schemes are adjointable? If there's a switch in
there then it may not work.

I might also try a Shapiro filter - is this adjointable? Also, do you have
any general tips/docs/example data files on the use of Shapiro filter?

Thanks

Dan