[MITgcm-support] Help me get rid of my noise!
Ryan Abernathey
ryan.abernathey at gmail.com
Wed May 7 15:57:26 EDT 2014
Thanks for all your suggestions.
I tried turning on convective adjustment with
ivdc_kappa=1.0,
in my data file. I also enabled
#define ALLOW_GGL90_SMOOTH
in GGL90_OPTIONS.h
It looks like the high values of W went down a little bit over the slope, a
slight improvement. So I guess I am satisfied that, as Sonia put it, the
solution is not "pretty" but is physically correct.
I am still worried about the "stripes" of high and low W within the mixed
layer between -60 and -50.
Thanks again,
Ryan
On Wed, May 7, 2014 at 3:14 PM, David Ferreira <dfer at mit.edu> wrote:
> Patrick's right, you need to have some convection with ggl90, but in
> case it's not enough: have you tried activating the smoother in ggl90 ?
> Might help too.
> david
>
>
>
> On 5/7/14 7:52 PM, Patrick Heimbach wrote:
>
> Ryan,
>
> yes, switching on convective adjustment is still advised in conjunction with ggl90 (in contrast to KPP where the non-local term Gamma should take care of this).
>
> Cheers
> p.
>
> On May 7, 2014, at 2:43 PM, Ryan Abernathey <ryan.abernathey at gmail.com> <ryan.abernathey at gmail.com> wrote:
>
>
> Thanks for all the helpful suggestions everyone!
> I am now convinced that this isn't as "unphysical" as I thought.
>
> In response to Christopher's suggestions, does anyone know whether I should have convective adjustment active together with GGL90? I know for KPP you are supposed to turn off all convective adjustment. Is that also the case for GGL?
>
> Thanks again,
> Ryan
>
>
> On Wed, May 7, 2014 at 2:10 PM, Christopher L. P. Wolfe <christopher.wolfe at stonybrook.edu> <christopher.wolfe at stonybrook.edu> wrote:
>
> My guess that you’re seeing hydrostatic convection. You don’t seem to have set either cAdjFreq or ivdc_kappa and my understanding is that they both default to zero. So, your model’s trying to do convection through the momentum equations and the convective scale is smaller than your grid spacing (hydrostatic convection runs to the grid scale).
>
> Christopher
>
>
> On May 7, 2014, at 1:20 PM, Ryan Abernathey <ryan.abernathey at gmail.com> <ryan.abernathey at gmail.com> wrote:
>
>
> I guess my image didn't make it through on that last email.
>
>
>
> On Wed, May 7, 2014 at 1:11 PM, Ryan Abernathey <ryan.abernathey at gmail.com> <ryan.abernathey at gmail.com> wrote:
> Hello MITgcm Support Group,
>
> I am running a 2D model (depth and latitude) in a Southern Ocean configuration. I am trying to eliminate what appears to be unphysical grid-scale noise in my W field--columns of alternately very high and low W. This is a snapshot of W after 5 years of spinup. This pattern is steady in time.
>
> As you can see, the noise is concentrated mainly (but not exclusively) over the shelf break. A possibly related issue is that my cg2d solver is relatively far from convergence:
> cg2d: Sum(rhs),rhsMax = -5.46186194609888E-05 6.64557830841732E-02
>
> My model is 0.5 degree resolution in latitude, meaning that it is relatively high resolution compared to the tutorials (e.g. ideal_2D_oce), but of course it has no eddies or any temporal variability at all, since it is 2D. This puts it in an unknown [at least to me] parameter regime.
>
> I have tried cranking the lateral viscosity (both laplacian and biharmonic) up as high as they will go. I have enabled both linear and quadratic bottom drag. I am NOT using the CD scheme; the noise was much worse with it enabled. I am NOT using KPP but rather GGL90. (I tried with KPP and it was not very different.)
>
> I would appreciate any suggestions you have about how to get rid of this noise and achieve smooth fields.
>
> Thanks,
> Ryan
>
> ####### data file ##########
> &PARM01
> useAreaViscLength=.FALSE.,
> viscAr=1.0E-03,
> viscA4=1e12,
> viscAh=1e5,
> tempAdvScheme=77,
> # KrT fields are read in from a file
> diffKrT=0.0E-5,
> diffKhT=0.0,
> diffK4T=0.0,
> staggerTimeStep=.TRUE.,
> # initial vertical profiles of T and S
> tRef=44*0.0,
> sRef=44*35.0,
> # equation of state
> eosType='LINEAR',
> rhonil=1035.,
> eosType='LINEAR',
> tAlpha=2.000000E-04,
> sBeta=0.000000E+00,
> saltStepping=.FALSE.,
> # boundary conditions
> no_slip_sides=.FALSE.,
> no_slip_bottom=.TRUE.,
> # needed for points above the U sponge layer (shelf)
> bottomDragLinear=1.E-3,
> bottomDragQuadratic = 0.0012
> # additional drag is done with RBCS
> # physical parameters
> gravity=9.810000E+00,
> # implicit diffusion and convective adjustment
> implicitDiffusion=.TRUE.,
> implicitViscosity=.TRUE.,
> implicitFreeSurface=.TRUE.,
> # exact volume conservation
> # exactConserv=.TRUE.,
> # C-V scheme for Coriolis term
> useCDscheme=.FALSE.,
> # partial cells for smooth topography
> hFacMin=5.000000E-02,
> # file IO stuff
> readBinaryPrec=64,
> useSingleCpuIO=.TRUE.,
> debugLevel=1,
> &
> # elliptic solver parameters
> &PARM02
> cg2dMaxIters=500,
> cg2dTargetResidual=1E-09,
> &
> # timestepping parameters
> &PARM03
> nIter0=0000172800,
> # nTimeSteps=1000,
> # nTimeSteps=12961,
> nTimeSteps=345600,
> # deltaT=3600.0,
> # deltaT=2400.,
> # deltaT=1800.0,
> deltaT=900.0,
> abEps=0.1,
> # needed for sea ice
> forcing_In_AB = .FALSE.,
> ChkptFreq=31104000.
> pChkptFreq=311040000.,
> taveFreq=311040000.0,
> dumpFreq=31104000.,
> monitorFreq=2592000.0,
> tauThetaClimRelax=2592000.0,
> dumpInitAndLast=.TRUE.,
> pickupStrictlyMatch=.FALSE.,
> &
> # gridding parameters
> &PARM04
> usingSphericalPolarGrid=.TRUE.,
> delX=0.5,
> delY=100*0.5,
> xgOrigin=0.,
> ygOrigin=-80.,
> delR=10., 10., 10., 10., 10., 10., 10., 12., 14.,
> 16., 19., 22., 26., 30., 36., 42., 50., 60.,
> 72., 85., 100., 120., 140., 166., 200., 200., 200.,
> 200., 200., 200., 200., 200., 200., 200., 200., 200.,
> 200., 200., 200., 200., 200., 200., 200., 200.
> &
> # Input datasets
> &PARM05
> bathyFile='bathyFile_shelf_2Dchan.bin',
> zonalWindFile='zonalWindFile.bin',
> thetaClimFile='thetaClimFile_WOA98_2Dchan.bin',
> diffKrFile='diffKrFile_nosponge_2Dchan.bin',
> hydrogThetaFile='hydrogThetaFile_WOA98_2Dchan.bin',
> hydrogSaltFile='hydrogSaltFile_WOA98_2Dchan.bin',
> &
>
> ########## data.pkg ##############
> &PACKAGES
> useMNC = .TRUE.,
> useGMREDI = .TRUE.,
> useGGL90 = .TRUE.,
> useRBCS = .TRUE.,
> &
>
> ######## data.gmredi ##############
> &GM_PARM01
> GM_MNC = .FALSE,
> GM_AdvForm = .TRUE.,
> GM_UseBVP = .TRUE.,
> GM_background_K = 2000,
> GM_isopycK = 2000,
> GM_Kmin_horiz = 50,
> GM_taper_scheme = 'dm95',
> GM_BVP_ModeNumber = 1,
> GM_BVP_cMin = .1,
> &end
>
> ########### data.ggl90 ############
> &GGL90_PARM01
> GGL90writeState=.FALSE.,
> GGL90TKEmin = 1.e-7,
> mxlMaxFlag=2,
> GGL90mixingLengthMin=3.,
> &
>
>
> ############ SIZE.h ##############
> ...
> PARAMETER (
> & sNx = 1,
> & sNy = 100,
> & OLx = 4,
> & OLy = 4,
> & nSx = 1,
> & nSy = 1,
> & nPx = 1,
> & nPy = 1,
> & Nx = sNx*nSx*nPx,
> & Ny = sNy*nSy*nPy,
> & Nr = 44)
> ...
>
>
> <W.png>_______________________________________________
>
> MITgcm-support mailing listMITgcm-support at mitgcm.orghttp://mitgcm.org/mailman/listinfo/mitgcm-support
>
>
> _______________________________________________
> MITgcm-support mailing listMITgcm-support at mitgcm.orghttp://mitgcm.org/mailman/listinfo/mitgcm-support
>
>
> _______________________________________________
> MITgcm-support mailing listMITgcm-support at mitgcm.orghttp://mitgcm.org/mailman/listinfo/mitgcm-support
>
>
> ---
> Patrick Heimbach | heimbach at mit.edu | http://www.mit.edu/~heimbach
> MIT | EAPS 54-1420 | 77 Massachusetts Ave | Cambridge MA 02139 USA
> FON +1-617-253-5259 | FAX +1-617-253-4464 | SKYPE patrick.heimbach
>
>
>
>
> _______________________________________________
> MITgcm-support mailing listMITgcm-support at mitgcm.orghttp://mitgcm.org/mailman/listinfo/mitgcm-support
>
>
>
> _______________________________________________
> 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/20140507/e8826c4b/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: W.png
Type: image/png
Size: 34113 bytes
Desc: not available
URL: <http://mitgcm.org/pipermail/mitgcm-support/attachments/20140507/e8826c4b/attachment-0001.png>
More information about the MITgcm-support
mailing list