[MITgcm-support] Need assistance for tuning currents in the physical model

[kunal madkaiker]

Hi Matt and Dimtris,

Yes, I have supplied U&V currents along with T and S along all the four
boundaries.
The above attached data.obcs file was for my simulation in which I had used
Orlanski. So I had to # the obcs inputs. Sorry for the confusion.
If you observe the plots, there isn't much difference between my closed
boundary o/p and the o/p for which I used the Orlanski scheme.

Matt, thanks for mentioning the 3rd and 4th point. I hadn't thought about
these aspects in my previous simulations. Will check more.

Kunal


On Thu, Jul 9, 2020 at 10:03 AM Matthew Mazloff <mmazloff at ucsd.edu> wrote:

> Hi Kunal
>
> Not sure about the plots, but I strongly suspect that you may have OBCS
> problems. Looks like you may have a waveguide along the boundary…
> Try plotting w variance and see if the open boundaries light up. If you
> have a wave guide and enhanced vertical exchange you will begin to change
> the stratification in your entire domain
>
> A few things
> 1) where did you get this data.obcs?
>  set  useOBCSsponge=.TRUE., and uncomment (and likely adjust)
> # Urelaxobcsinner=172800.E0,
> # Urelaxobcsbound=172800.E0,
> # Vrelaxobcsinner=172800.E0,
> # Vrelaxobcsbound=172800.E0,
> # spongeThickness=04,
> I don’t have experience with orlanski…not sure about it.
>
> 2) You are not prescribing any T, S, U, V on the boundary? That is what
> your data.obcs shows….if that is true that is why your currents look
> poor. The open boundaries are vital to get correct for your domain as oscar
> shows you have a strong through flow
>
> 3) you really should make the bathymetry flat in the open boundary
> restoring region. Otherwise prescribing normal velocity is equivalent to
> prescribing vertical velocity as well, via divergence, and you don’t want
> spurious vertical velocity
>
> 4) I prefer to set tangential velocity to zero on the boundary, otherwise
> you do need to prescribe a divergence somewhere….Also setting tangential =
> 0 damps waves
>
> -Matt
>
>
> On Jul 8, 2020, at 9:06 PM, Dimitris Menemenlis <dmenemenlis at gmail.com>
> wrote:
>
> Hi Kunal, impressive start!
>
> It’s best if you keep discussion in MITgcm-support, however,
> that way other MITgcm users might contribute to or benefit from the
> discussion.
>
> So you have no river runoff?  Could that explain the lack of low salinity
> region south of Gulf of Khambhat?
>
> And you are not specifying currents at the open boundaries?  Could that
> explain, in part, the differences with OSCAR?
>
> I have no explanation for the warm SST bias.  Which bulk formulae are you
> using?
> Or maybe there is a problem in your surface boundary conditions, e.g.,
> wrong time of year?
> Or could you have chosen an anomalously warm year to simulate?
> Remember that WOA is a really weird multi-year space-time average.
> So comparing one month of your model simulation with WOA will not
> necessarily match.
>
> D.
>
>
> On Jul 8, 2020, at 8:45 PM, kunal madkaiker <kunal.madkaiker02 at gmail.com>
> wrote:
>
> Hello Matt and Dimitris
>
> Thanks a lot to you both for introducing me to MITgcm_contrib. I am sorry
> for not knowing it since I have just started my modeling research work with
> MITgcm.
> I would like to share with you my model simulation output plot for
> temperature and salinity. Kindly let me know your thoughts.
> Simulated current profile isnt coming good, as expected.
>
> Also, attaching my data files and SIZE.h. Thanks once again.
>
> Kunal
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.mitgcm.org/pipermail/mitgcm-support/attachments/20200709/637f3116/attachment-0001.html>