[MITgcm-support] questions re. external forcing

[Holly Dail]

Thanks for the idea David.  Since my configuration uses dynamically  
calculated FW and latent heat transport associated with evaporation,  
balancing the model input files by themselves may not be enough.

For anyone who may want an answer to these questions as well, I've  
talked to a few people offline and now have these impressions:
- Trying to mix gregorian calendar input files with the 'model'  
calendar is a bad idea.  If using a realistic, modern setting, using  
Gregorian everywhere probably makes the most sense.  In my case, I'll  
probably develop simplified forcing fields with a 360 day year.

- There seems to be no built-in way to do balanced fluxes over longer  
periods than one time step, so I'll modify the code to do it.  The sea  
ice package complicates the process somewhat ... its important to  
calculate the mis-balance after the sea ice package has calculated  
fluxes in sea ice regions, and to consider whether to consider shifts  
to and from sea ice as losses from the ocean or not (in my case  
probably not as I could then 'correct' for lost freshwater when I  
indeed want salinity to increase w/ sea ice formation).

thanks -
Holly

On Mar 27, 2009, at Mar 27 , 1:44 PM, David Wang wrote:

> Hi Holly,
>
> Concerning flux balancing. There could be better solutions. But what  
> I did is to go back to the model input files, calculate the annual-,  
> global (global in terms of your entire model domain) mean and  
> subtract this number (hopefully not a large one) from the original  
> flux at every grid point. Then feed the model with the "balanced"  
> flux input data. This preserves the seasonal cycle.
>
> HTH,
> D.
>
> On Thu, Mar 26, 2009 at 7:17 PM, Holly Dail <hdail at mit.edu> wrote:
> Hello -
>
> I have a couple of questions regarding external forcing and would be  
> most grateful for advice.  For background, my model setup is:
> - 1 degree isotropic N. Atlantic with closed boundaries near 20 S  
> and 70 N (well, closed for now anyway)
> - external forcing with NCEP fields & bulk formula
> - interested in quasi steady-state behavior over hundreds of years  
> (e.g. tracer distributions)
> - I want to include seasons, but plan to neglect (for the moment)  
> inter-annual variability
>
> On to the questions:
> - I'd like to use the model calendar for my simulation / output, and  
> Gregorian for the EXF fields.  When one sets  TheCalendar='model' in  
> data.cal, what does the model assume about the calendar for the  
> fields defined in data.exf?  Are 365 days/year fields interpolated  
> to 360 days, or do the seasons 'move' by 5 days/year of simulation  
> due to calendar mis-match?
>
> - I'd like to have balanced net heat and freshwater fluxes on an  
> annual basis to hopefully reach a quasi steady-state.   I tried  
> adding #define ALLOW_BALANCE_FLUXES in CPP_OPTIONS.h and  
> balanceQnet=.TRUE., balanceEmPmR=.TRUE. in data.  This works great,  
> but I believe sets the net fluxes to zero in each time step.  This  
> is problematic for seasonality.   Since evaporation is computed on  
> the fly, I am also limited in what I can do in 'correcting' heat  
> flux and freshwater fields before the run.
> In the absence of other options, I'm thinking of modifying the code  
> to compute the net heat flux and FW flux accumulated over a year and  
> use it to 'correct' the following year's fluxes.    Are there  
> better / cleaner ways to do this?
>
> Thanks for any help!
> Holly
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support
>
>
>
> -- 
> life grows, death doesn't. there is no dress rehearsal.
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support