[MITgcm-support] Using a sponge layer to relax temperature

Tue Mar 3 17:16:12 EST 2015

Hi Jody.

Actually, I don’t think I can combine them. I have already rewritten the rbcs_add_tendency.F file to force my plume in a specific way, and it’s in a way that I do not wish to relax my temperature at the boundaries. So I am fairly certain combining the plume temperature and the relaxation temperature in the relaxTFile won’t work.

Best,
Ashley
On Mar 3, 2015, at 5:02 PM, Stroman, Ashley <acs08g at my.fsu.edu<mailto:acs08g at my.fsu.edu>> wrote:

Hi Jody.

That is very true. I actually never thought about combining them. Thanks for bringing that up. I definitely will consider this.

Best,
Ashley
On Mar 3, 2015, at 4:56 PM, Jody Klymak <jklymak at uvic.ca<mailto:jklymak at uvic.ca>> wrote:

Hi Ashley,

I’m probably not following your problem.  Your relaxTFile can have temperatures that go to initial conditions at the boundaries, and go to your plume values where your plume is.  They aren’t mutually exclusive.  And if you want the time scales to be different, the relaxMaskTFile can specify numbers from 1 to 0 that allow you to adjust the time scales differently everywhere.

Cheers,   Jody

On 3 Mar 2015, at  13:50 PM, Stroman, Ashley <acs08g at my.fsu.edu<mailto:acs08g at my.fsu.edu>> wrote:

Hi Jody.

My issue is that I am already using the rbcs to force temperature in the center of my domain to create a plume, but I also want to use the rbcs package to relax temperature on the boundaries to the initial conditions. So the hacks that Dustin is discussing is to allow more than one relaxation temperature field. The routine modified by Sciascia allows multiple momentum relaxation fields, but not for temperature.

I think it is doable to rewrite Sciascia’s code to also include multiple temperature relaxation fields, but I’m not sure if it’s worth it when I can just use a sponge layer to relax temperature on the boundaries.

Ashley
On Mar 3, 2015, at 4:28 PM, Jody Klymak <jklymak at uvic.ca<mailto:jklymak at uvic.ca>> wrote:

Hi Ashley,

rbcs is pretty flexible.  Here is my data.rbcs for runs where I was doing tidal forcing.  I didn’t need separate masks for U,V and T, but you might if you are are forcing your plume as temperature only and letting U and V freely evolve.

I don’t know about the hacks Dustin is discussing, but its hard to see why they are necessary as rbcs is so flexible.  For me the only disadvantage for big runs is having to specify the fields everywhere for every rbcsForcingPeriod.

Cheers,   Jody

# RBCS package parameters:
 &RBCS_PARM01
#------------------------------------------------------------------------------
# switches
#------------------------------------------------------------------------------
   useRBCtemp=.TRUE.,
   useRBCvVel=.TRUE.,
   useRBCuVel=.TRUE.,
#   useRBCsalt=.TRUE.,
#------------------------------------------------------------------------------
#- relaxation times
# This is the timescale where "sponge=1".
#------------------------------------------------------------------------------
    tauRelaxU=120000.,
    tauRelaxV=120000.,
    tauRelaxT=120000.,
#------------------------------------------------------------------------------
# masks - #1 and #2 = T,S , #3 = tracers ;
# masks  U,V have explicit name - if left out then TMask = used
#------------------------------------------------------------------------------
# Probably need different U and V masks because they are on a slightly different
# grid
   relaxMaskFile='spongeweight.bin',
#   relaxMaskUFile='usponge.bin',
#   relaxMaskVFile='vsponge.bin',

#------------------------------------------------------------------------------
# files containing relaxation flds  have explicit name
#------------------------------------------------------------------------------
   relaxTFile='Tforce.bin',
   relaxUFile='Uforce.bin',
   relaxVFile='Vforce.bin',
#------------------------------------------------------------------------------
   rbcsIniter=0,
# 1/24 tidal cycle
   rbcsForcingPeriod=1860.,
   rbcsForcingCycle=44640.,
/

On 3 Mar 2015, at  11:28 AM, Stroman, Ashley <acs08g at my.fsu.edu<mailto:acs08g at my.fsu.edu>> wrote:

Hi Dustin.

After looking through the code, it appears that the modified rbcs package from Siascia is only for multiple relaxation fields regarding momentum and not temperature. So if I wanted to use this in my simulations, then I’d have to rewrite the routine to also include temperature, which might be doable. Not sure if it’s easiest to rewrite the rbcs code to include multiple relaxations fields for temperature or just use the OBCS package to use a sponge layer though.

Best,
Ashley
On Mar 2, 2015, at 1:44 PM, Dustin Carroll <dcarroll at uoregon.edu<mailto:dcarroll at uoregon.edu>> wrote:

Hi Ashley,

Ryan has good advice here. if you need to relax your plume and boundary temperature at different timescales, Roberta Sciascia added a nice mod. to RBCS pkg under MITgcm_contrib/sciascia/rbcs that allows for multiple RBCS fields with independent relaxation times. You can then have multiple masks / relaxation files (i.e., one for your plume and one for the boundaries).

Dustin

On Mar 2, 2015, at 10:31 AM, Stroman, Ashley <acs08g at my.fsu.edu<mailto:acs08g at my.fsu.edu>> wrote:

Well, I already use the RBCS package to force a plume in the middle of the domain, so I didn’t try using the RBCS package for the temperature on the boundaries because I wasn’t sure if you could use the RBCS package twice in one simulation.

-Ashley
On Mar 2, 2015, at 1:28 PM, Ryan Abernathey <ryan.abernathey at gmail.com<mailto:ryan.abernathey at gmail.com>> wrote:

Just wanted to point out, as a possible alternative, the RBCS package.
http://mitgcm.org/public/r2_manual/latest/online_documents/node237.html

RBCS allows you to relax to model fields to some prescribed field (with a variable relaxation rate) throughout the whole domain. It is somewhat simpler to use and understand than OBCS.

If you don't need actual open boundary features (e.g. Orlanski, inflow / outflow), RBCS is more than adequate for making sponge layers.

-Ryan

On Mon, Mar 2, 2015 at 12:56 PM, Stroman, Ashley <acs08g at my.fsu.edu<mailto:acs08g at my.fsu.edu>> wrote:
Hi Dustin.

I have two specific questions regarding my problem:

1.) Is it possible to use Orlanski for momentum on the boundaries, while also using a sponge layer to restore temperature to the initial conditions on the boundaries? I’m not sure if both are allowed, and I’ve had no luck finding this answer online.

2.) What is the difference between useOBCSsponge and OBCSsponge_Theta in the data.obcs file? My thought is that the latter is specifically used for theta while the former is not, but I am not 100% positive. There’s not much information on the latter parameter except that it turns on the theta sponge along the boundaries, so I’m not sure if the parameter is used, 1) what it specifically means, and 2) what is additionally required.

Thanks.
Ashley
On Feb 27, 2015, at 8:06 PM, Dustin Carroll <spanishcastlemagik at gmail.com<mailto:spanishcastlemagik at gmail.com>> wrote:

Hi Ashley,

Can you provide a bit more information on the problem you are having? I find that the more specific you are, the easier it is for people to help on this support group.

Section 6.3.1 in the docs (see below) describes how to set up your data.obcs file, also look at verification/exp4 for an example w/ open boundaries.
http://mitgcm.org/public/r2_manual/latest/online_documents/node236.html

If you want to restore your boundaries to the prescribed ICs then you’ll need to generate binary files for each boundary that is specified in data.obcs,

Cheers,
Dustin

On Feb 27, 2015, at 11:56 AM, Stroman, Ashley <acs08g at my.fsu.edu<mailto:acs08g at my.fsu.edu>> wrote:

Hello.

I am wanting to use the OBCS package to implement a sponge layer on the boundaries to relax temperature to the initial conditions. I am having trouble finding what is needed in terms of the data.obcs file as well as any any additional binary files for temperature that are required. Can someone shed some light on this? Thanks.

Best,
Ashley Stroman
_______________________________________________
MITgcm-support mailing list
MITgcm-support at mitgcm.org<mailto:MITgcm-support at mitgcm.org>
http://mitgcm.org/mailman/listinfo/mitgcm-support

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

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

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

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

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

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

--
Jody Klymak
http://web.uvic.ca/~jklymak/

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

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

--
Jody Klymak
http://web.uvic.ca/~jklymak/

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

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

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mitgcm.org/pipermail/mitgcm-support/attachments/20150303/6bbd43bb/attachment-0001.htm>