[MITgcm-support] waves in a rectangular basin!

[Matthew Robert Mazloff]

Hi Van,

For model stability, the Courant number, Co = U dt/dx, should be less than one.
You have to determine how the velocity scales in your problem.  If you want to
look at surface gravity waves, we find in your configuration that the wave
speed = sqrt(gH)~ 20m/s.  With a dt of 10 and a dx of 40 you have 
Co ~ 5 which is too large.

Reducing the Courant number would be a good place to start.

-Matt


Quoting Van Thinh Nguyen <vtnguyen at moisie.math.uwaterloo.ca>:

> Hi Ed,
> 
> I try to use MITgcm to simulate a surface waves (also for a internal
> waves!) in a close rectangular box (x*y*h=4000m*4000m*40m). This model is
> based on "exp1" of MITgcm's tutorials, the domain is closed around by
> solid walls, and only has some changes (you will see on data file attached.
> 
> My computation was crashed many many times when the surface wave hit the
> right wall then reflected from there to hit the wave from the left due to
> wind force. I even tried to decreased: visAh, visAz, time step, ...
> I got the following warning:
> 
> -------------------
> **WARNNING** CONFIG_CHECK: useMNC is TRUE and #undef ALLOW_MNC
> SOLUTION IS HEADING OUT OF BOUNDS: tMin,tMax=  -4.161E+02   9.254E+02
> MON_SOLUTION: STOPPING CALCULATION
> STOP MON_SOLUTION: STOPPED DUE TO EXTREME VALUES OF SOLUTION statement
> executed
> -------------------
> 
> #data file
> ------------------
> # Model parameters
> # Continuous equation parameters
>  &PARM01
>  tRef=  20.0 , 20.0 , 20.0 , 20.0 , 20.0 ,
>         20.0 , 20.0 , 20.0 , 20.0 , 20.0 ,
>         19.99 , 19.99 , 19.99 , 19.98 , 19.98 ,
>         19.96 , 19.95 , 19.92 , 19.88 , 19.82 ,
>         19.73 , 19.61 , 19.43 , 19.17 , 18.81 ,
>         18.32 , 17.69 , 16.90 , 15.99 , 15.00 ,
>         14.01 , 13.10 , 12.31 , 11.68 , 11.19 ,
>         10.83 , 10.57 , 10.39 , 10.27 , 10.18 ,
>  sRef=  10., 10., 10., 10., 10.,
>         10., 10., 10., 10., 10.,
>         10., 10., 10., 10., 10.,
>         10., 10., 10., 10., 10.,
>         10., 10., 10., 10., 10.,
>         10., 10., 10., 10., 10.,
>         10., 10., 10., 10., 10.,
>         10., 10., 10., 10., 10.,
>  viscAz=1.E-6,
>  viscAh=1.E-6,
>  no_slip_sides=.TRUE.,
>  no_slip_bottom=.TRUE.,
>  diffKhT=1.E-6,
>  diffKzT=1.E-6,
>  beta=1.E-11,
>  tAlpha=2.E-4,
>  sBeta =0.,
>  gravity=9.81,
>  rigidLid=.FALSE.,
>  implicitFreeSurface=.TRUE.,
>  nonHydrostatic=.TRUE.,
>  eosType='LINEAR',
>  readBinaryPrec=64,
>  &
> # Elliptic solver parameters
>  &PARM02
>  cg2dMaxIters=1000,
>  cg2dTargetResidual=1.E-8,
>  &
> # Time stepping parameters
>  &PARM03
>  startTime=0.,
> #endTime=311040000,
>  endTime=432000.0,
>  deltaTmom=10.0,
>  deltaTtracer=10.0,
>  abEps=0.1,
>  pChkptFreq=1200.0,
>  chkptFreq=1200.0,
>  dumpFreq=1200.0
> # dumpFreq=2592000.0,
>  monitorFreq=1.,
>  &
> # Gridding parameters
>  &PARM04
>  usingCartesianGrid=.TRUE.,
>  usingSphericalPolarGrid=.FALSE.,
>  delX=100*40,
>  delY=100*40,
>  delZ=  1., 1., 1., 1., 1.,
>         1., 1., 1., 1., 1.,
>         1., 1., 1., 1., 1.,
>         1., 1., 1., 1., 1.,
>         1., 1., 1., 1., 1.,
>         1., 1., 1., 1., 1.,
>         1., 1., 1., 1., 1.,
>         1., 1., 1., 1., 1.,
>  &
>  &PARM05
>  bathyFile='topog.box',
>  hydrogThetaFile=,
>  hydrogSaltFile=,
>  zonalWindFile='windx.sin_y',
>  meridWindFile=,
>  &
> --------------------------
> 
> #geometry from gendata.m file
> -------------------------
> ieee='b';
> accuracy='real*8';
> 
> Ho=40;
> nx=100;
> ny=100;
> 
> % Flat bottom at z=-Ho
> h=-Ho*ones(nx,ny);
> % Walls
> h(1,:)=0;
> h(:,1)=0;
> h(end,:)=0;
> h(:,end)=0;
> fid=fopen('topog.box','w',ieee); fwrite(fid,h,accuracy); fclose(fid);
> 
> % Wind-stress
> tauMax=0.1;
> x=((1:nx)-0.5)/(nx-1); % nx-1 accounts for a solid wall
> y=((1:ny)-0.5)/(ny-1); % ny-1 accounts for a solid wall
> [X,Y]=ndgrid(x,y);
> tau=tauMax*sin(pi*Y);
> fid=fopen('windx.sin_y','w',ieee); fwrite(fid,tau,accuracy); fclose(fid);
> 
> ------------------
> 
> Could you please take a look and may have some mentions about this?
> 
> Thanks so much for your helps.
> 
> Van Thinh
> 
> ************************************
> Van Thinh Nguyen, PhD. Postdoc fellow
> Dept. of Applied Mathematics
> University of Waterloo, Canada
> Email: vtnguyen at uwaterloo.ca
> *************************************
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support
>