[MITgcm-devel] seaice adjoint and EVP

[Jinlun Zhang]

Your log plot is quite innovative. I doubt coding bug would cause this. 
This is because a bug is unlikely to be a function of timestep. I would 
assume that the numerical elastic waves have high frequencies. When 
timestep gets smaller, the EVP solution is supposed to approach VP 
solution and the unphysical waves would be damped out. But what you show 
is that one has to use very small timestep (close to that for the fully 
explicit method), and even so there is still residual noisiness. Is it 
possible to run a case with timestep of 0.2 s? I am curious about how 
close EVP can get to VP(LSR).
Jinlun

Dimitris Menemenlis wrote:

> one other way to emphasize differences is to plot log10(1-AREA):
> http://ecco2.jpl.nasa.gov/data1/arctic/output/tests/figs/maskedAREA4320.ps 
>
>
> while the LSR solution is very smooth, the EVP solution, even with 1-s 
> time steps, has some noise
>
> question is whether this noise is intrinsic to EVP scheme or whether 
> it's a residual bug in the code?
>