[MITgcm-support] hfac question

[David Hebert]

Hi Matt,

Fortunately, we do have depth one point interior equal to the depth on 
the boundary. So it sounds like following the depth rule you outlined  
below it is better to use hFacC on the all boundaries.


Thanks for your help,

David

On 10/06/09 11:37, Matthew Mazloff wrote:
> Hi David,
>
> Yeah, using hFacC for transport on the open boundaries seems 
> consistent with my understanding of the implementation.  I think you 
> are correct that the value of  hFacW came from periodicity -- but is 
> not relevant to the prescription of OBCS values, and this is why hFacC 
> is used.
>
> I believe this must be why when one uses OBCS they should have 
> Depth(1,:) = Depth(2,:).  So that hFacW(2,:) = hFacC(1,:) and 
> everything remains consistent.
>
> -Matt
>
>
>
>
> On Oct 6, 2009, at 9:28 AM, David Hebert wrote:
>
>> Matt,
>>
>> Along those lines, I notice in obcs_calc.F, for the balance routines, 
>> that hFacC's are used instead of hFacW for E/W boundary and hFacS for 
>> N/S boundaries. Could that also be related to the hFacW and hfacC 
>> assuming periodic BC/s?
>>
>> Thanks,
>>
>> David
>>
>> On 10/06/09 11:09, Matthew Mazloff wrote:
>>> Hi David,
>>>
>>> It is a bit odd that it gave you 90 for hFacW(1,:), but I guess it 
>>> is fine as I believe what happens on the western open boundary is 
>>> that it sets U(1,:) and U(2,:) to the prescribed value and sets 
>>> (assumes) zero divergence right on the boundary center (helps 
>>> pressure solver).  So yes, I believe what is truly coming into the 
>>> dynamics -- and what is entering the domain -- is sum(U(2,:).*drF .* 
>>> hfacW(2,:)).
>>>
>>> -Matt
>>>
>>>
>>> On Oct 6, 2009, at 8:56 AM, David Hebert wrote:
>>>
>>>> Hi Matt,
>>>>
>>>> The point I mentioned is a W boundary point (should  have noted 
>>>> that before, sorry). All surrounding points  are  also 100m. I do 
>>>> note that the Eastern boundary is 90m. So if hFacW will obtain 
>>>> depth on Eastern bounary is there something different I should do 
>>>> on the West? i.e., seems I should use hFacW(2,:)?
>>>>
>>>> Thanks
>>>>
>>>> David
>>>>
>>>> On 10/06/09 10:49, Matthew Mazloff wrote:
>>>>> Hi David,
>>>>>
>>>>> This is correct.  You must have sloping topography and the depth 
>>>>> at the west side of the cell is 90m, while its 100m in the 
>>>>> center.  U is calculated on the west face, so the total volume 
>>>>> transport is
>>>>> sum(U(1,:).*drF .* hfacW(1,:)).
>>>>>
>>>>> Likewise one should use hFacS for V transport.
>>>>>
>>>>> -Matt
>>>>>
>>>>>
>>>>> On Oct 6, 2009, at 8:38 AM, David Hebert wrote:
>>>>>
>>>>>> Hi all,
>>>>>>
>>>>>> I am attempting to balance barotropic inflow/outflow on the 
>>>>>> boundaries of a 2D simulation. I ran the model one timestep to 
>>>>>> obtain the hFac's in order to more accurately determine the depth 
>>>>>> the model uses. The result from Depth.data at one point is 100m.
>>>>>>
>>>>>> In order to balance U, according to manual, I would use hFacW. I 
>>>>>> get a very different result if I do sum(drF * hFacW). Results 
>>>>>> from matlab (for 2D hFac's)...
>>>>>>
>>>>>> sum(drF .* hfacW(1,:)) = 90
>>>>>> sum(drF .* hfacC(1,:)) = 100
>>>>>>
>>>>>> Seems I don't quite understand hfacW. Any ideas on what I did 
>>>>>> wrong here?
>>>>>>
>>>>>> Thanks
>>>>>>
>>>>>> David
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> MITgcm-support mailing list
>>>>>> MITgcm-support at mitgcm.org
>>>>>> http://mitgcm.org/mailman/listinfo/mitgcm-support
>>
>> _______________________________________________
>> MITgcm-support mailing list
>> MITgcm-support at mitgcm.org
>> http://mitgcm.org/mailman/listinfo/mitgcm-support
>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support

-- 
David A. Hebert
NRC Postdoc
Naval Research Lab
Stennis Space Center, MS 39529
david.hebert.ctr at nrlssc.navy.mil
Phone: (228) 688-5846
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mitgcm.org/pipermail/mitgcm-support/attachments/20091006/909eec3b/attachment.htm>