[MITgcm-support] Restart simulation with biogeochemical packages

Mon Nov 23 08:37:06 EST 2020

Dear Jonathan and Jean-Michel,

Thanks a lot for your help!

Just to clarify the setup I am using, I run coupled atmosphere-ocean-thice simulations for studying the Permian-Triassic climate.
I have run such simulations without DIC until statistical equilibrium and now I am adding the geochemical module.

Following your suggestion, in my new tests I have not set input files in data.dic, and I have used the default values for silicate concentration for the moment. Moreover, I have used initial files for tracers concentrations and it seems to work well, thanks again!

At present I am using TRACERS_Iter0 = nIter0 in data.ptracers. I saw I can use TRACERS_Iter0 > nIter0 instead, in order to activate the geochemical module at a following time with respect to the dynamical pickup (temp, salt…) iteration. However, I am not able to set TRACERS_Iter0 < nIter0 as in cfc_example, because the model exits with an error due to the absence of pickup_ptracer.nIter0 file.

Kind regards,

Charline

> Hi Jonathan and Charline,
>
> Thanks Jonathan for providing these detailed informations.
>
> Just would like to mention one feature of pkg/ptracers (used for bio-geo chemical tracers)
> that might be relevant here:
> You can start all the tracers (from each tracer, one tracer initial concentration set from file
>   PTRACERS_initialFile(#) = 'File_Name', ) while the dynamics (Temp, Salinity and flow-field)
> is re-started from pickup files. This is done with ptracer parameter "PTRACERS_Iter0".
> And there are few examples in verification experiment that do use it, such as:
>   cfc_example/input/data.ptracers
>
> Cheers,
> Jean-Michel
>
> On Thu, Nov 05, 2020 at 07:49:57PM +0000, Jonathan Lauderdale wrote:
>> Hi Charline,
>>
>> Without knowing exactly what your set up is, your approach sounds reasonable. You can also initialize the physical model from binary files so you don???t need to start entirely from rest/scratch (see for example the input file list for ???data??? file here:https://github.com/MITgcm/MITgcm/blob/master/model/src/ini_parms.F#L286).
>>
>> The input files in data.dic (ice, wind, silica, atmospheric pressure) are for prescribed forcing of pkg/dic (i.e. not something that requires pickups). ???DIC_iceFile" is prescribed sea ice fraction that affects light and gas fluxes, ???DIC_windFile" is for prescribed wind speed that also affects gas transfer, same with pressure. The frequency and periodicity of the forcing is controlled by either the main model forcing cycle or by setting ???DIC_forcingPeriod??? or "DIC_forcingCycle???. If you???re using other packages that compute these things (such as pkg/seaice, or atmospheric model), then they are probably overwritten in dic_fields_update.F. However, DIC_silicaFile is different, because it???s not computed elsewhere or carried as a tracer in pkg/dic. Nevertheless, the silicate concentration is initialized to a constant value if you don???t supply anything and that is then used in the pH/pCO2 routines. You may need to consider if this is ok or what value(s) fit your needs bette
>   r.
>> Indeed, if PTRACERS_initialFile is not set, then your tracer fields are initialized to the values set in PTRACERS_ref, your tracer concentrations will be zero.
>>
>> Certain tracers that have a freshwater influence are sensitive to PTRACERS_EvPrRn like DIC and ALK, so yes, setting those (to zero excludes any dissolved concentrations in rainwater/evaporate) allows dilution/concentration. Others you may want to enforce conservation (PO4/DOP) so leaving PTRACERS_EvPrRn unset does no FW forcing for that tracer. Your values for PTRACERS_ref may also affect concentration/dilution here, depending on your options for freshwater forcing and surface boundary conditions (see ptracers_forcing_surf.F): you might find the situation (per ???convertFW2Salt??? value or ???useRealFreshWaterFlux??? flag) where the dilution/concentration effect is the difference between PTRACERS_ref at the surface and PTRACERS_EvPrRn i.e. if they???re both zero, then there???s no FW induced effect, where you might want one.
>>
>> If you???re still getting weird pH values, then you might want to consider using the Munhoven (2013) carbonate system solver (#define CARBONCHEM_SOLVESAPHE in DIC_OPTIONS.h and selectPHsolver>0 in data.dic)???it???s probably more complicated and less efficient that the Follows et al (2006) scheme, but it may be a bit more robust to initialization shocks and varied paleoclimate conditions.
>>
>> Cheers,
>>
>> Jonathan
>> ___________________________________________________________________________
>> Dr. Jonathan M. Lauderdale
>> Research Scientist
>> Department of Earth, Atmosphere and Planetary Sciences
>> Massachusetts Institute of Technology
>> 77 Massachusetts Avenue (54-1518)
>> Cambridge, MA 02139, USA
>> Office: +1 617 324 1568
>> Cell   : +1 617 304 5661
>> Email:mailto:jml1 at mit.edu  <mailto:jml1 at mit.edu>
>> Web:http://paocweb.mit.edu/people/jml1  <http://paocweb.mit.edu/people/jml1>
>> Twitter:https://twitter.com/jon_lauderdale
>>   <https://twitter.com/jon_lauderdale>Blog:https://seamanticscience.wordpress.com/  <https://seamanticscience.wordpress.com/>
>> Git:https://github.com/seamanticscience  <https://github.com/seamanticscience>
>> ___________________________________________________________________________
>>
>>> On Nov 5, 2020, at 12:45 PM, Charline Ragon<Charline.Ragon at unige.ch>  wrote:
>>>
>>> Dear all,
>>>
>>> I would like to restart a simulation originally performed without the biogeochemical packages by activating them (thus including PTRACERS, GCHEM and DIC). Since there are no some of the needed pickups (namely, pickup_cpl and pickup_aimCo2), I have created them by performing a simulation that starts from nIter0=0. My idea is to use such pickups to restart my simulation, is this the right procedure?
>>>
>>> Doing like this, I obtain reasonable values for the tracers if I activate DIC_BOUNDS in order to avoid out-of-range pH values.
>>>
>>> Since I am performing paleo-climate simulations with huge uncertainties on the concentrations of the different tracers, I am wondering:
>>>
>>> 1. Is it correct to set PTRACERS_ref to zero values and start from PTRACERS_Iter0=nIter0 without any initial file? (if I understand well, all fields would be initialised to 0 in this case)
>>> 2. Is the model very sensitive to PTRACERS_EvPrRn? Or is it only necessary to give a certain value for allowing dilution/concentration? Which value?
>>> 3. Which is the role of the input files in data.dic (ice, wind, silica, atmospheric pressure)? Should they replace pickup files?
>>>
>>> Thanks in advance for any suggestion!
>>>
>>> Best,
>>> Charline
>>>
>>>
>>> --
>>> Charline RAGON
>>> PhD Student
>>> Group of Applied Physics & Institute for Environmental Sciences
>>> University of Geneva, Switzerland
>>> _______________________________________________
>>> MITgcm-support mailing list
>>> MITgcm-support at mitgcm.org
>>> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support
>
>> _______________________________________________
>> MITgcm-support mailing list
>> MITgcm-support at mitgcm.org
>> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support

--
Charline RAGON
PhD Student
Group of Applied Physics & Institute for Environmental Sciences
University of Geneva, Switzerland

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.mitgcm.org/pipermail/mitgcm-support/attachments/20201123/2244ef0a/attachment.html>