[MITgcm-support] Ocean currents veering due to bathymetry

[kunal madkaiker]

Hi Matt and Martin,

Thank you both for your suggestions. I ran the model with T,S,U and V at
boundary forcing as suggested by Matt.
Also as recommended by Martin, changed hFacMin =0.3 from 0.9, and set
 OBCSfixTopo=.TRUE. from FALSE. Also trying to improve
the grid simultaneously.

And that fixed the issue. Please refer to the figure attached. I also gave
a run adding heat and moisture fluxes from NCEP reanalysis.
But it looks like Model SST is overestimating by ~1.25 degC when compared
with WOA18 SST. Is it because fluxes are being fed in as
monthly climatology and not real time daily values? I simulated the model
for 30 days.
Thanks again!

Regards
Kunal

------------
From: kunal madkaiker <kunal.madkaiker02 at gmail.com>
Date: Fri, Jul 24, 2020 at 9:52 AM
To: MITgcm Support <mitgcm-support at mitgcm.org>

Dear MITgcm users,

I am trying to simulate ocean currents along the West coast of India in a
high resolution grid (360x780) for the month of September. For now, I am
providing monthly climatology of August Temperature and Salinity as initial
forcing along with September zonal and meridional wind stress. Currently,
using the Orlanski scheme for obcs and no boundary forcings. Model is
simulated for 10 days.

But the currents are not coming correct. The West Indian Coastal Current
(WICC) which flows equatorward (during Sept) isn't being captured at all.
After plotting the bathymetry, I observed that the currents are being
veered northward by the Arabian sea continental shelf (Please refer to
Picture1.png attached (Day 04-10)). Is that possible? My bathymetry bin
file is just a 2D mesh with depth values. Also, my coastal currents are
very weak.

Attaching my data,data.obcs and data.exf file. Please let me know if any
additional information is required from my end.
Kindly advise. All suggestions are welcomed.

Regards
Kunal


----------
From: Matthew Mazloff <mmazloff at ucsd.edu>
Date: Fri, Jul 24, 2020 at 10:12 AM
To: <mitgcm-support at mitgcm.org>


1) You are setting U and V = 0 as an initial condition? Then likely U and V
aren’t spun up yet after just 10 days.

2) regional models are very sensitive to obcs. You aren’t prescribing a
UVTS on the boundary? That will kill your coastal currents.

Sounds like the problems are just with your (lack of) inputs

Matt >
<data.obcs><bathymetry.png><data><data.exf><Picture1.png>_______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
>
https://urldefense.com/v3/__http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support__;!!Mih3wA!WJ7uZPG1M2dThC8r1pphYDlMIcAf9vYRKuuJejYIZeUqSlQmMjjY1ew3p3guXOG0Ug$

_______________________________________________
MITgcm-support mailing list
MITgcm-support at mitgcm.org
http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support


----------
From: Martin Losch <Martin.Losch at awi.de>
Date: Fri, Jul 24, 2020 at 3:09 PM
To: MITgcm Support <mitgcm-support at mitgcm.org>


Hi Kunal,

I have no idea of the loca hydrography and circulation and don’t know what
to expect, but there are few things that may be problematic in your setup:

- check the forcing: the surface stresses are aligned with the grid
direction, i.e. ustress is not directed towards the east, but towards
increasing model index “i” (which is sometimes called “east” in the code).
So your wind stress needs to be rotated according to your grid orientation
before it is read by the model. Probably you have already thought about this
- with your obcs settings there is no net flow through the open boundaries.
Is that what you want? If the current system is only local, that this is OK
but a quick google search tells me that the southward WICC should actually
leave your domain. You’d need some sort of boundary conditions for that. I
would suggest that you look into the Stevens boundary conditions (described
in the documenation) where you can specify a barotropic flow field normal
to the boundaries.
- a 10-day simulation period may be very short for any system to develop
- there’s no buoyancy forcing. Is that important or do you only want wind
driven circulation?

small things that you may want to adjust:
- your viscosities are very high
- your horizontal grid aspect ratio is very high (dx/dy = 2891.9791/939.661
= 3), that could cause numerical problems (e.g. CFL problems or stability
problems with viscosity). If you can still change that ratio to 1, I would
do that
- for delZ/delR the rule of thumb is that delZ(k+1)/delZ(k) < 1.4, i.e.
there shouldn’t be too large jumps in the grid spacing so that the
numerical approximations are not compromised
- hFacMin = 0.9 is close to turning off the partial cells. I’d use
something like 0.3 or even 0.1 if possible. with your 5m grid spacing this
may be problematic but you can set minimum thickness with
hFacMinDz/hFacMinDr (=5 or 10)

- OBCSfixTopo=.FALSE., is not a good choice, use the default (.True.)

Hope that helps,
Martin >
<data.obcs><bathymetry.png><data><data.exf><Picture1.png>_______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.mitgcm.org/pipermail/mitgcm-support/attachments/20200729/59d7eefa/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: SST_compare.gif
Type: image/gif
Size: 95666 bytes
Desc: not available
URL: <http://mailman.mitgcm.org/pipermail/mitgcm-support/attachments/20200729/59d7eefa/attachment-0001.gif>