[MITgcm-support] Warning and time stepping issue
Anthony Coletti
ajcolett at geo.umass.edu
Thu Jan 16 11:14:31 EST 2014
Hi Christoph and Martin,
Thank you for your responses!
@Martin: I am looking through the manual now to see if I can tweak my run to make it run faster (5000s time step is unbearably slow for such course resolution).
@Christoph: My ice thicknesses are not out of the ordinary at all. I think the thickest amount of ice is about 4 meters thick.
Anthony
Anthony J. Coletti
Climate System Research Center
Department of Geosciences
Morrill Building
611 N. Pleasant Street
233 Morrill Science Center
University of Massachusetts-Amherst
paleoclimate.org
Email: ajcolett at geo.umass.edu
http://blogs.umass.edu/ajcolett/
“For me, I am driven by two main philosophies: know more today about the world than I knew yesterday and lessen the suffering of others. You'd be surprised how far that gets you.” ― Neil deGrasse Tyson
On Jan 16, 2014, at 10:59 AM, Christoph Voelker <christoph.voelker at awi.de> wrote:
> Hi Anthony and Martin,
>
> I suspect thta it is not the CFL criterium, but the sea-ice in
> conjunction with the nonlinear free surface, under extreme cases of ice
> formation. I had a similar isue in some last glacial maximum model runs
> that I did, where the free surface code basically complained that the
> submerged part of the sea ice got too thick. When that happened I also
> got very large negative eta values. There was a not particulaly elegant
> trick around it that I don't remember at the moment, I'll check. But
> Anthony, maybe you can check your ice thickness?
>
> Cheers, Christoph
>
> On 1/16/14 4:41 PM, Martin Losch wrote:
>> Hi Anthony,
>>
>> I usually find the stability “analysis” in the tutorials quite illustrative. For example, for the 4deg global ocean experiment <http://mitgcm.org/public/r2_manual/latest/online_documents/node129.html>, you can find a couple of stability criteria: CFL for advection and viscosity, but also for inertial oscillations. The Coriolis is treated explicitly and I believe, that this limits your time step for the momentum equations. The only way around it is to use asynchronous time stepping (at the usual cost of distorted physics).
>>
>> Martin
>>
>> On Jan 16, 2014, at 10:13 AM, Anthony Coletti <ajcolett at geo.umass.edu> wrote:
>>
>>> I should be more specific with my problem. Apologies.
>>>
>>> See below for the original problem:
>>>
>>> I am using the sea-ice package and external forcing packages (along with other default packages) to do a control LGM simulation of the Arctic. I am currently trying to see how much sea ice I can grow in the Arctic during this period. I have prescribed most of the atmosphere, the initial SST and Salinity fields.
>>>
>>> My data file looks like this:
>>>
>>> # ====================
>>> # | Model parameters |
>>> # ====================
>>> #
>>> # Continuous equation parameters
>>> &PARM01
>>> tRef=15*20.,
>>> sRef=15*35.,
>>> viscAh =3.E5,
>>> viscAr =1.E-3,
>>> diffKhT=0.,
>>> diffK4T=0.,
>>> diffKrT=3.E-5,
>>> diffKhS=0.,
>>> diffK4S=0.,
>>> diffKrS=3.E-5,
>>> ivdc_kappa=10.,
>>> implicitDiffusion=.TRUE.,
>>> gravity=9.81,
>>> rhoConst=1035.,
>>> rhoConstFresh=1000.,
>>> eosType='JMD95Z',
>>> staggerTimeStep=.TRUE.,
>>> vectorInvariantMomentum=.TRUE.,
>>> implicitFreeSurface=.TRUE.,
>>> exactConserv=.TRUE.,
>>> select_rStar=2,
>>> nonlinFreeSurf=4,
>>> hFacInf=0.2,
>>> hFacSup=2.0,
>>> useRealFreshWaterFlux=.TRUE.,
>>> hFacMin=.1,
>>> hFacMinDr=20.,
>>> #- to check that it has no impact:
>>> # doResetHFactors=.TRUE.,
>>> # tempAdvScheme=7,
>>> # saltAdvScheme=7,
>>> readBinaryPrec=32,
>>> writeBinaryPrec=32,
>>> useSingleCpuIO=.TRUE.,
>>> &
>>>
>>> # Elliptic solver parameters
>>> &PARM02
>>> cg2dMaxIters=200,
>>> #cg2dTargetResidual=1.E-9,
>>> cg2dTargetResWunit=1.E-14,
>>> &
>>>
>>> # Time stepping parameters
>>> &PARM03
>>> nIter0=0.,
>>> endTime=31557600000.,
>>> deltaT= 1500.,
>>> abEps = 0.1,
>>> forcing_In_AB=.FALSE.,
>>> pickupStrictlyMatch=.FALSE.,
>>> cAdjFreq=0.,
>>> pChkptFreq = 31557600.,
>>> chkptFreq = 15778800.,
>>> monitorFreq = 21600.,
>>> &
>>>
>>> # Gridding parameters
>>> &PARM04
>>> usingCurvilinearGrid=.TRUE.,
>>> horizGridFile='../../../input.files/grid_inputs/grid_cs32',
>>> delR= 50., 70., 100., 140., 190.,
>>> 240., 290., 340., 390., 440.,
>>> 490., 540., 590., 640., 690.,
>>> &
>>>
>>> # Input datasets
>>> &PARM05
>>> bathyFile ='../../../input.files/LGM_cntr/lgm+ish_cntr_bathy.bin',
>>> hydrogThetaFile='../../../input.files/LGM_cntr/sfc_T_coldstart_32x32x15.bin',
>>> hydrogSaltFile ='../../../input.files/LGM_cntr/oce_salt_coldstart_32x32x15.bin',
>>> &
>>>
>>>
>>> When I lower my time step BELOW 5000. s, it does not crash and is stable the entire run. I also have bathymetry set to LGM levels (adding 120m to the bathymetry files) to make the ocean more shallow.
>>>
>>>
>>> Anthony
>>>
>>>
>>> Anthony J. Coletti
>>> Climate System Research Center
>>> Department of Geosciences
>>> Morrill Building
>>> 611 N. Pleasant Street
>>> 233 Morrill Science Center
>>> University of Massachusetts-Amherst
>>> paleoclimate.org
>>> Email: ajcolett at geo.umass.edu
>>> http://blogs.umass.edu/ajcolett/
>>>
>>> “For me, I am driven by two main philosophies: know more today about the world than I knew yesterday and lessen the suffering of others. You'd be surprised how far that gets you.” ― Neil deGrasse Tyson
>>>
>>>
>>>
>>>
>>> On Jan 16, 2014, at 9:45 AM, Anthony Coletti <ajcolett at geo.umass.edu> wrote:
>>>
>>>> Hi all,
>>>>
>>>> I am trying to run a course 32x32x15 (6 faces) run to get some preliminary data. Unfortunately, my time step will not exceed 5000.s. ~ Two minutes into the run, it crashes and I get the corresponding error::
>>>>
>>>> WARNING: r*FacC > hFacSup at 1 pts : bi,bj,Thid,Iter= 4 1 1 106
>>>> e.g. at i,j= 2 1 ; rStarFac,H,eta = 2.028435 2.200000E+01 2.262556E+01
>>>> WARNING: r*FacC > hFacSup at 1 pts : bi,bj,Thid,Iter= 4 1 1 107
>>>> e.g. at i,j= 2 1 ; rStarFac,H,eta = 2.333596 2.200000E+01 2.933912E+01
>>>> WARNING: r*FacW < hFacInf at 2 pts : bi,bj,Thid,Iter= 2 1 1 108
>>>> e.g. at i,j= 5 2 ; rStarFac,H,eta = 0.094709 1.368000E+03 -6.740896E+02 -1.803211E+03
>>>> STOP in CALC_R_STAR : too SMALL rStarFacW !
>>>> WARNING: r*FacS < hFacInf at 1 pts : bi,bj,Thid,Iter= 2 1 1 108
>>>> e.g. at i,j= 3 2 ; rStarFac,H,eta = -0.186928 2.400000E+02 5.703576E+01 -6.277609E+02
>>>> STOP in CALC_R_STAR : too SMALL rStarFacS !
>>>> WARNING: r*FacC < hFacInf at 2 pts : bi,bj,Thid,Iter= 2 1 1 108
>>>> e.g. at i,j= 5 2 ; rStarFac,H,eta = -0.152948 1.564000E+03 -1.803211E+03
>>>> STOP in CALC_R_STAR : too SMALL rStarFacC !
>>>>
>>>> If you need any other info, please let me know. Any help would be appreciated!
>>>>
>>>>
>>>> Thanks!
>>>>
>>>> Anthony
>>>>
>>>> Anthony J. Coletti
>>>> Climate System Research Center
>>>> Department of Geosciences
>>>> Morrill Building
>>>> 611 N. Pleasant Street
>>>> 233 Morrill Science Center
>>>> University of Massachusetts-Amherst
>>>> paleoclimate.org
>>>> Email: ajcolett at geo.umass.edu
>>>> http://blogs.umass.edu/ajcolett/
>>>>
>>>> “For me, I am driven by two main philosophies: know more today about the world than I knew yesterday and lessen the suffering of others. You'd be surprised how far that gets you.” ― Neil deGrasse Tyson
>>>>
>>>>
>>>>
>>>>
>>> _______________________________________________
>>> 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
>
>
> --
> Christoph Voelker
> Alfred Wegener Institute for Polar and Marine Research
> Am Handelshafen 12
> 27570 Bremerhaven, Germany
> e: Christoph.Voelker at awi.de
> t: +49 471 4831 1848
>
>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mitgcm.org/pipermail/mitgcm-support/attachments/20140116/609fc7bd/attachment-0001.htm>
More information about the MITgcm-support
mailing list