[MITgcm-support] SOM instability after checkpoint 64e

[Martin Losch]

Hi Spencer,

in doc/tag-index I found that there were changes to the advection packages in checkpoint64e. Maybe you can diff the files (e.g. like this: <https://github.com/MITgcm/MITgcm/commit/d077f4a7023faea230432dc180400e4970aff0ff#diff-5cf10394edefefcf714f07bf62cf6d40> or using other git capabilities) and figure out which change causes the in instability in your setup. 

Martin

> On 21. Aug 2018, at 21:23, Spencer Jones <spencerj at ldeo.columbia.edu> wrote:
> 
> Hi everyone, 
> I'm trying to update my code tree before I start a new project, but I'm finding that SOM is unstable in my setup from checkpoint 64e onwards. With the flux limiter off (advection scheme 80), the temperature becomes large and negative. With the flux limiter on (advection scheme 81), the temperature remains positive, but I get an unphysical-looking standing wave in the zonal direction at mid-latitudes. 
> 
> I have tried changing the viscosity, diffusivity and various other things but to no avail. Does anyone know what's going on? My data file is pasted below. 
> Thanks very much, 
> Spencer 
> 
> # ====================
> # | Model parameters |
> # ====================
> #
> # Continuous equation parameters
>  &PARM01
>  tRef = 32*0.,
>  sRef = 32*35.,
>  no_slip_bottom=.TRUE.,
>  no_slip_sides=.TRUE.,
>  viscAr=1.E-4,
>  viscAh=40.E3,
> # diffKrT=1.E-5,
>  diffKrNrT=0.0083, 0.0065, 0.0038,0.0014, 0.00039,0.000092,
>            0.00003, 0.00002,0.00002,0.00002,0.00002,0.00002,
>            0.00002,0.00002,0.00002,0.00002,0.00002,0.00002,0.00002,0.00002,
>            0.00002,0.00002,0.00002,0.00002,0.00002,0.00002,0.00002,0.00002,
>            0.00002,0.00002,0.00002,0.00002,
> # diffKrS=2.E-5,
>  diffKrNrS=0.0083, 0.0065, 0.0038,0.0014, 0.00039,0.000092,
>            0.00003, 0.00002,0.00002,0.00002,0.00002,0.00002,
>            0.00002,0.00002,0.00002,0.00002,0.00002,0.00002,0.00002,0.00002,
>            0.00002,0.00002,0.00002,0.00002,0.00002,0.00002,0.00002,0.00002,
>            0.00002,0.00002,0.00002,0.00002,
>  rhonil=1035.,
>  rhoConstFresh=1000.,
>  eosType = 'LINEAR',
>  ivdc_kappa=10.,
>  implicitDiffusion=.TRUE., 
>  bottomDragLinear=1.E-3,
>  exactConserv=.TRUE.,
>  useRealFreshWaterFlux=.FALSE.,
>  useSingleCpuIO=.TRUE.,
> # turn on looped cells
>  hFacMin=.1,
>  hFacMindr=50.,
> # set precision of data files
>  readBinaryPrec=32,
>  tempAdvScheme=81,
>  saltAdvScheme=81,
>  staggerTimeStep=.TRUE.,
>  debugLevel=-1, 
>  &
> 
> # Elliptic solver parameters
>  &PARM02
>  cg2dMaxIters=500,
>  cg2dTargetResidual=1.E-13,
>  printResidualFreq=37606140,
>  &
> 
> # Time stepping parameters
>  &PARM03
>  nIter0=13000000,
>  nTimeSteps = 4800000,
> # 100 years of integration will yield a reasonable flow field
> # startTime =          0.,
> # endTime   = 3110400000.,
>  deltaTmom = 900.,
>  deltaTtracer= 10800.,
>  deltaTClock = 10800.,
>  deltaTfreesurf= 5400.,
>  abEps = 0.1,
>  pChkptFreq= 4320000000.,
>  chkptFreq=  4320000000.,
>  dumpFreq= 43200000000000.,
>  monitorFreq=86400000.,
> # 10 days restoring timescale for temperature
>  tauThetaClimRelax=  864000.,
> # 6 months restoring timescale for salinity
> # tauSaltClimRelax = 15552000.,
> # periodicExternalForcing=.FALSE.,
> # externForcingPeriod=2592000.,
> # externForcingCycle=31104000.,
>  pickupStrictlyMatch=.FALSE., 
>  &
> 
> # Gridding parameters
>  &PARM04
>  usingSphericalPolarGrid=.TRUE.,
>  delR= 32*125.,
>  ygOrigin=-70.,
>  dySpacing=0.9722,
>  dxSpacing=0.9722,
>  &
> 
> # Input datasets
>  &PARM05
>  bathyFile=      'bathy.bin',
>  zonalWindFile=  'taux.bin',
>  meridWindFile=  'tauy.bin',
>  thetaClimFile=  'sst2.bin',
> # saltClimFile=   'lev_sss.bin',
> # surfQFile=      'ncep_qnet.bin',
> # the_run_name=   'global_oce_latlon',
> # fresh water flux is turned on, comment next line to it turn off
> # (maybe better with surface salinity restoring)
>  EmPmRFile=      'sflux.bin',
>  &
> 
> 
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mailman.mitgcm.org/mailman/listinfo/mitgcm-support