[MITgcm-support] Bulk Force and SSS restoring
Amato Evan
ate9c at virginia.edu
Wed Feb 17 14:57:04 EST 2010
Hi Martin,
Thanks again for your time, I really appreciate it!
> Amato,
> to be honest, I have not fully understand the problem, but that aside:
I seem to be really good at confusing people!
>
> I meant "global_with_exf/input.yearly" (I was not aware that there
> are two sub-input directories in this experiments), so
> theCalendar="model", is an option and implies 12 30-day-months
Hmmm... I don't remember why I didn't try with this setup before. OK,
this is great to know, thanks!
> Now for your problem: As far as I can tell you use the
> "external_fields_load"-method to load the forcing, right? (with
> these filenames:
> & zonalWindFile, meridWindFile,
> & thetaClimFile, saltClimFile,
> & surfQnetFile, surfQswFile, EmPmRfile
> in data, PARM05) for your spin-up. Your qnet = qsw(net) + lw(net) +
> qsh + qlh. lw(net) = lwup - lwdown, where
> qlwup=emissivity*boltzmann*SST^4 or so. Now, if you want to change
> your long wave downward radiation, why not compute qnet-lwdown +
> new_lwdown. The lwdown is part of the NCEP fields (at least the
> variant that I once downloaded from the LDEO ingrid server (http://iridl.ldeo.columbia.edu/
> , specifically: <http://iridl.ldeo.columbia.edu/SOURCES/.NOAA/.NCEP/.EMC/.CMB/.T40/
> >) and that is the basis of most of my climatological simulations
> without sea-ice.
I actually am doing this in a current model run which uses neither EXF
nor BULK_FORCE.
Here I am differencing two runs, one where I force with reanalysis
Qnet, and one where I force with reanalysis Qnet + lw & sw aerosol
forcing. The problem with calculating Qnet offline in this way is that
lwup is esentially fixed at climatology and is not responsive to (in
this case mostly) the SST cooling forced by the reduction of downward
sw radiation by aerosols. This ends up being too liberal an estimate
of the temperature response to aerosol forcing.
> In exf, you can specify wind (or stress), lwdown, swdown, humidity,
> surface air temperature, precipitation (and with an additional flag,
> also evaporation), or net lwflux and net swflux (depends on the CPP-
> flags, see EXF_OPTIONS.h for more details).
This is what I am doing (u & v stress, lw down, sw down, air q &
temp). I'd like to leave off precip and just restore to SSS since this
is how I spun up the model to begin with (using "external fields load").
>
> Martin
>
> On Feb 17, 2010, at 5:07 PM, Amato Evan wrote:
>
>> Hi Martin,
>>
>> Thanks a lot for chiming in here!
>>
>>> in data.cal you can specify the 'model calendar' (see
>>> global_with_exf for an example), so that should not stop you from
>>> use exf instead of bulk_force. Although my experience with
>>> bulk_force is very limited, the exf does have quite a few options,
>>> and I think it should do what you need.
>> I would certainly much rather use EXF. global_with_exf has
>> "TheCalendar='gregorian'". I did incorporate cal and exf into my
>> current model setup, but I could not get the model to run if I had
>> "TheCalendar='model'" (it did work w/gregorian though). And then I
>> saw something on the mailing list where someone said that with EXF
>> you have to use the gregorian calendar. Then I stopped playing with
>> EXF since I don't want to deal w/leap years!
>>
>>> Did you have a look at global_ocean.cs32x15/input.thsice/ for an
>>> example use of bulk_force?
>> I did, I can get BULK_FORCE going in my model setup, I'm just
>> trying to determine if I can use it w/o the model calculating
>> freshwater fluxes from a rain or runoff file (only SSS restoring),
>> which I couldn't deduce from the code/documentation. Not such a
>> huge thing if not I suppose. I'm just trying to save myself some
>> time.
>>
>> Sincerely,
>> Amato
>>
>>
>>
>>>
>>> Martin
>>> On Feb 17, 2010, at 3:58 PM, Amato Evan wrote:
>>>
>>>> Hi Stefano!
>>>>
>>>> I was hoping to use the model calendar (just b/c it's more
>>>> straightforward) and as far as I can tell, if one uses EXF then
>>>> calendar has to be set to "Gregorian". Is this right?
>>>>
>>>> I should say that another reason for only restoring to SSS is
>>>> that this is how I have spun up the model, so I'd like to stay
>>>> consistent on this front as well.
>>>>
>>>> Thanks for your time, I really appreciate it!
>>>>
>>>> Sincerely,
>>>> Amato
>>>>
>>>> On Feb 17, 2010, at 4:37 AM, Stefano Querin wrote:
>>>>
>>>>> Hi Amato,
>>>>>
>>>>> just a few ideas...
>>>>>
>>>>> Have you ever tried to use the EXF package for your simulations?
>>>>> It
>>>>> should be more complete and flexible than the BULK_FORCE
>>>>> package. You
>>>>> can also restore SSS to prescribed values with it.
>>>>>
>>>>> I don't think that the computation of freshwater fluxes using
>>>>> runoff
>>>>> and precipitation input files makes the model run that slow...
>>>>> (but
>>>>> I'm just relying on my experience, I never made any test regarding
>>>>> this). On the contrary, if you meant that the creation and
>>>>> handling
>>>>> (offline) of runoff and precip files is somehow cumbersome and
>>>>> slow,
>>>>> then SSS restoring should be a more desirable choice.
>>>>>
>>>>> Anyway, the EXF package has many available options.
>>>>>
>>>>> Hope this helps...
>>>>>
>>>>> Cheers!
>>>>>
>>>>> S.
>>>>>
>>>>>
>>>>>
>>>>> On 16 Feb 2010, at 19:50:23, Amato Evan wrote:
>>>>>
>>>>>> Hello Everyone,
>>>>>>
>>>>>> I am trying to estimate the effect of surface forcing by
>>>>>> volcanic-
>>>>>> aerosols on tropical SST. To do this I have a 1x1 degree model
>>>>>> setup
>>>>>> w/23 vertical levels. After spining up the model for 130 years
>>>>>> using
>>>>>> NCEP and WOA data (and SSS restoring), I would like to "turn-
>>>>>> on" the
>>>>>> BULK_FORCE package in order to change downward lw & sw fluxes in
>>>>>> accordance with volcanic aerosol surface forcing (something I
>>>>>> calculate offline w/satellite obs. and an RT model).
>>>>>>
>>>>>> I am giving the BULK_FORCE pkg wind stress values (calcWindStress
>>>>>> = .FALSE.), but I would like to run w/surface salinity restoring,
>>>>>> and not have the model calculate freshwater fluxes using runoff
>>>>>> and
>>>>>> precip input files. My motivation is just to make the model runs
>>>>>> complete more quickly. What I can't determine from the
>>>>>> documentation
>>>>>> and source code is if I leave:
>>>>>>
>>>>>> RainFile= ' ',
>>>>>> RunoffFile= ' ',
>>>>>> EmPFile= ' ',
>>>>>>
>>>>>> in data.blk, will bulk_force just restore to surface salinity
>>>>>> on the
>>>>>> time scales I define in "data"? If anyone has experience with
>>>>>> this,
>>>>>> I would really appreciate input. Also, the contents of build/,
>>>>>> code/, and input/ for my setup can be found here:
>>>>>>
>>>>>> http://trane.evsc.virginia.edu/mitgcm/
>>>>>>
>>>>>> If anyone is so inclined to look them over.
>>>>>>
>>>>>> Thanks!
>>>>>>
>>>>>> Sincerely,
>>>>>> Amato
>>
>> ---
>> Amato Evan
>> Asst. Professor
>> Dept. of Environmental Sciences
>> University of Virginia
>> 434.243.7711
>> aevan at virginia.edu
>> http://trane.evsc.virginia.edu
>>
>>
>> _______________________________________________
>> 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
---
Amato Evan
Asst. Professor
Dept. of Environmental Sciences
University of Virginia
434.243.7711
aevan at virginia.edu
http://trane.evsc.virginia.edu
More information about the MITgcm-support
mailing list