[MITgcm-support] [EXTERNAL] Inquiry about deltaTtracer, deltaTmom, and deltaTClock in MITgcm

Martin Losch Martin.Losch at awi.de
Wed Sep 4 01:39:09 EDT 2024


Please follow Dimitris’ recommendation. 

Asynchronous timestepping has been used for long spinup simulations with coarse resolution.

The different timestep parameters do *not* mean that the model is stepped deltaTclock/deltaTtracer times, etc. Instead it means that say at time step two, the model time witll be 2*deltaTclock (used to determine when new forcing is to be read, etc), and the tracer equations will have been stepped 2 times with deltaTtracer, i.e. the tendencies are very small, when deltaTtracer is very small, and this means that it will not develop very much, e.g 0.002 sec. The seaice models use deltaTtracer as the default time step, unless you specify is explicitly, so the same applies.

Martin

> On 4. Sep 2024, at 03:51, Menemenlis, Dimitris (US 329B) <dimitris.menemenlis at jpl.nasa.gov> wrote:
> 
> from MITgcm/model/inc/PARAMS.h
> 
> C     deltaT    :: Default timestep ( s )
> C     deltaTClock  :: Timestep used as model "clock". This determines the
> C                    IO frequencies and is used in tagging output. It can
> C                    be totally different to the dynamical time. Typically
> C                    it will be the deep-water timestep for accelerated runs.
> C                    Frequency of checkpointing and dumping of the model state
> C                    are referenced to this clock. ( s )
> C     deltaTMom    :: Timestep for momemtum equations ( s )
> C     dTtracerLev  :: Timestep for tracer equations ( s ), function of level k
> 
> but I recommend that for your problem you just set deltaT so that you have:
> 
> deltaTClock = deltaTMom = dTracelLev = deltaT
> 
> D.
> 
> 
>> On Sep 3, 2024, at 6:05 PM, 唐鸿 <tanghong22 at mails.ucas.ac.cn> wrote:
>> 
>> Dear MITgcm Support Team,
>> 
>> 
>> 
>> I hope this message finds you well.
>> 
>> 
>> 
>> I am currently working on a simulation project using the MITgcm to study the dynamics of a lake, including the freezing process during winter. I have encountered some confusion regarding the behavior of the model with respect to the parameters deltaTtracer, deltaTmom, and deltaTClock, and I am seeking your guidance.
>> 
>> 
>> 
>> In my understanding, the model should operate as follows when given the settings:
>> 
>> deltaTClock = 3600s
>> deltaTtracer = 1800s
>> deltaTmom = 100s
>> I believed that over one deltaTClock period, the tracer equations would be updated twice, and the momentum equations would be updated 36 times.
>> 
>> 
>> 
>> However, I have conducted two different simulations with the following configurations and results:
>> 
>> First simulation configuration:
>> 
>> nIter0 = 0
>> nTimeSteps = 13199
>> deltaTClock = 3600s
>> deltaTmom = 0.00001s
>> deltaTtracer = 3600s
>> The resulting ice_iceH output seemed more reasonable, with ice formation as expected.
>> Second simulation configuration:
>> 
>> nIter0 = 0
>> nTimeSteps = 13199
>> deltaTClock = 3600s
>> deltaTmom = 0.001s
>> deltaTtracer = 1s
>> Surprisingly, this simulation did not produce any ice formation.
>> The driving data spans from 2015-06-15 to 2016-12-15 with a time resolution of 3600 seconds. Intuitively, I think that the simulation results should be the same for both configurations , with only a difference in runtime. I am puzzled by the behavior of these parameters, particularly deltaTtracer and deltaTmom, and how they actually take effect within the model.
>> 
>> 
>> 
>> Could you please clarify how these parameters are supposed to function within the model, and whether my understanding aligns with the model's implementation?
>> 
>> 
>> 
>> 
>> I would greatly appreciate any suggestions or guidance you could provide. I am really looking forward to your valuable insights and suggestions.
>> 
>> 
>> 
>> Best regards,
>> 
>> 
>> 
>> Hong Tang
>> 
>> University of Chinese Academy of Chinese
>> 
>> 
>> 
>> _______________________________________________
>> MITgcm-support mailing list
>> MITgcm-support at mitgcm.org
>> https://urldefense.us/v3/__http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support__;!!PvBDto6Hs4WbVuu7!LSlrwpz8oaCFIoCLDNwhjzGA34lnifqINuFGNjrdRdGDEqdemAMz_vs9yiHkw7rxbo82eJI18RaVvkGRv9jg8nt862IEofGTkUJ8bnE$
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.mitgcm.org/pipermail/mitgcm-support/attachments/20240904/8190ddc5/attachment.html>


More information about the MITgcm-support mailing list