POINT=           14  MAX ETADOT VERTICAL VEL.=   3.671716816008357E-006

LON  =    136.000000000000       degrees

LAT  =    81.0519404509364       degrees

MAX V WIND=   298.250361785114

MAX U WIND=   550.836883184825

U WIND =   550.836883184825       IS TOO STRONG, EXPLOSION.

LEVEL=           91  POINT=            3

LON  =    162.000000000000       degrees

LAT  =    52.9823117777130       degrees

ABORT!   31  !U WIND TOO STRONG, EXPLOSION!!!

I'm curious as to what causes this problem. I've tried to run the model in debug mode and traceback but I cannot locate the cause. Has anyone seen this error before and has some experience with it?

4 Comments

  1. Glenn Carver

    Hi Marston,

    The most likely explanation is that the timestep is too long for the model resolution.

    What's the resolution and the timestep value?  For timestep, check the NODE.001.01 file and find the line similar to:

    SUDYN: TSTEP_TRAJ,TSTEP=   600.00000000000000        600.00000000000000

    The second number refers to the model timestep as set in the namelist.

    A few lines down you will see a line similar to:

    PTSTEP_OPT (optimal timestep computed in SUTSTEP) 0.36000000E+04

    This is an approximate value for the maximum timestep for a particular resolution. It's a guide only and often too high but it gives a ballpark figure.

    Let me know what you are using.

      Regards,   Glenn

     

  2. Marston Ward

    Hi Glenn,

    My timestep and model resolution are good. The model ran perfectly the day before and then now even after I recomple, it blows up. Here's snippet from NODE.001_01:

     

     ---- Set up model dynamics --------------------------------------------

     SUDYN: TSTEP_TRAJ,TSTEP=   2700.00000000000        2700.00000000000     

     SUDYN: NSMAX_TRAJ,NSMAX=         255         255

     SUDYN: NSMAX_BACKGR00, NSMAX_BACKGR01, NSMAX_BACKGR02, NSMAX_BACKGR03=

             255         255         255         255

     TESTVAR_MIX: NTESTVAR=           0

     TESTVAR_MIX: LCVTEST, LINTEST, LTLTEST, LADTEST, LGRTEST=  F F F F F

     TESTVAR_MIX: LTESTINC= F

     TESTVAR_MIX: LTWINTRUTH=  F , LTWINEXP=  F

     TESTVAR_MIX: TWINDEPART=  0.000000000000000E+000

     PTSTEP_OPT (optimal timestep computed in SUTSTEP) 0.36000000E+04

     WARNING: LNEWHD=.F. => RRDXTAU IGNORED (IF SET)

    ....

    MODULE YOMDYN

     * SUDYN: Timestep.

     TSTEP  = 0.27000000E+04 TDT    = 0.27000000E+04

     

    Any other likely cause?

  3. Marston Ward

    Hi Glenn,

    I found the problem. I forgot to remove some data pextra data. Since there was not allocated arrays OpenIFS generated a lot of
    SMILAG TRAJECTORY UNDERGROUND errors and the winds exploded. Now there are two ways to create that kind of error (smile)

    Appreciate the help,
    /M

    1. Glenn Carver

      Hi Marston,

      Glad you found it. Sounds like there was some memory overwriting happening.

        Glenn