Blowing up

[abeudin]

Hello,

I am currently trying to get familiar with Roms. I have set up a test case of my road stead with a simple bathymetry and analytical initial and boundary conditions as:

/*Simple configuration of my so-called "test case r2b"*/

#define r2b

#ifdef r2b

/*Model physics*/
#undef SOLVE3D
#define UV_QDRAG

/*Define computational grid and initial conditions*/
#define MASKING
#define ANA_INITIAL

/*Define boundary conditions*/
#define ANA_SMFLUX
#define ANA_FSOBC
#define FSOBC_REDUCED
#define WEST_FSCHAPMAN
#define WEST_M2REDUCED

/*Others options*/
#undef WET_DRY

#endif

For an obscure reason my run stops at the iteration n°3466 (over 43000 asked):

Blowing-up: Saving latest model state into RESTART file

WRT_RST - wrote re-start fields (Index=1) into time record = 0000003

I made sure that I could never had any dry point setting a bathymetry greater than 4.5m (owning an tidal amplitude of 1m).

I don't quite see where this problem comes from.

I would very much appreciated any suggestions.

Best regards.

-Alexis

[jcwarner]

"Blowing-up: Saving latest model state into RESTART file

WRT_RST - wrote re-start fields (Index=1) into time record = 0000003"

did u look at the rst file, time index #3???

[kate]

I believe that ROMS checks for velocity magnitude, not negative depths. In fact, I don't think ROMS will crash on negative depths (unless you are growing frazil ice there). The blow up check is happening in diag.F and you can add more verbose output there to tell how and where the trouble is. Also, a view of the rst file in ncview might show you something interesting.

[abeudin]

1) Indeed the blow-up is due to exceeding speed value:

maxspeed vs max_speed 20.0083756171520 20.0000000000000


2) When I looked at my "_rst.nc" file (compared to my "_his.nc" file) I found two peculiar points:

OCEAN_TIME
time since initialization 1:3 ... ... ...
(invalid coordinate axis)
ZETA free-surface 1:118 1:78 1:3 ...
UBAR vertically integrated u-momentu 1:117 1:78 1:3 ...
VBAR vertically integrated v-momentu 1:118 1:77 1:3 ...


-I have 3 vertical layers (cf 3rd column where it is written 1:3) while I am dealing with a 2D case (cf "_his.nc" file hereafter:

ZETA free-surface 1:118 1:78 ... 1:3466
UBAR vertically integrated u-momentu 1:117 1:78 ... 1:3466
VBAR vertically integrated v-momentu 1:118 1:77 ... 1:3466


-My ocean-time is odd.

I am grateful to you for pointing me out how to detect errors.
I hope to come back soon with this solved.

Regards.

-Alexis

[abeudin]

Hello,

I haven't figure out what was wrong yet. To make sure it was not an issue with strong slops and/or wetting and drying areas, I set my bathy to 50 meters everywhere in my domain.
I still get an 'blowing out' error message due to exceeding current value.
In my "rst.nc" file I get something strange regarding the following variables:


OCEAN_TIME
time since initialization 1:3 ... ... ...
(invalid coordinate axis)
ZETA free-surface 1:118 1:78 1:3 ...
UBAR vertically integrated u-momentu 1:117 1:78 1:3 ...
VBAR vertically integrated v-momentu 1:118 1:77 1:3 ...

Note: the four columns are respectively: X-axis,Y-axis,Z-axis,Time-axis


Why do I get 3 layers for zeta,ubar and zbar owning that I have set an 2D configuration?
What should be the dimension of my ocean_time (function of the number of nodes or a simple scalar)?


If somebody has encountered this issue or could redirect me, to more efficiently ruminate on this, I would be grateful to her/him.

Best regards.

-Alexis

[jcwarner]

"Why do I get 3 layers for zeta,ubar and zbar owning that I have set an 2D configuration?"
the 3 layers are the time levels.
did you plot the rst fields ? you can use a free ware like ncview to quickly see what is going on.