Strange wind stress behavior

[fragoso]

Dear all,

I´m running ROMS 3.0 for southwestern Atlantic Ocean using ECCO data (every 10 days) for initial and boundary conditions and NCEP winds (every 6 hours) for atmospheric forcing for a 2007 hindcast. I had to interrupt the run two times and performed a restart at day 140 and day 200.

I found some strange results like a very very strong upwelling and after checking the wind stress fields I found that during long periods the model considered a constant svstr and almost constant sustr.

I checked the input netcdf file and didn´t find anything wrong with the variables (apparently...). So, I think ROMS is considering constant values after some time, but I don´t know why. And after I restarted the model, the values were normal again!

The attached figure shows svstr and sustr for a given point for the netcdf input file and for the ROMS output (station file). I´m using a MacMini with gfortran (serial run).

Could someone give me a hint on what could be happening?

Thanks!

Regards,
Mauricio Fragoso
PROOCEANO - Brasil

[jcwarner]

can you look at the standard out information, and see if roms is doing the get_data steps?

[fragoso]

I think you got the point. The output (see below) shows that last stress reading was on day 273 - 06:00. At 12:00 it should have read the next field, but it didn´t until the end of the run. The ECCO fields are being normally read though. Do you know why this is happening?

Thank you.

Mauricio Fragoso
PROOCEANO - Brasil


65580 273 06:00:00 1.793244E-02 1.987897E+04 1.987899E+04 1.111513E+16
GET_2DFLD - surface u-momentum stress, t = 273 00:00:00
(File: roms_brasil_floats_frc_NCEP_Y2007.nc, Rec=1099, Index=1)
(Tmin= 0.0000 Tmax= 364.2500)
(Min = -6.38252306E-05 Max = 1.12004149E-04)
GET_2DFLD - surface v-momentum stress, t = 273 00:00:00
(File: roms_brasil_floats_frc_NCEP_Y2007.nc, Rec=1099, Index=1)
(Tmin= 0.0000 Tmax= 364.2500)
(Min = -3.43084050E-05 Max = 3.22088187E-04)
65581 273 06:06:00 1.793286E-02 1.987897E+04 1.987899E+04 1.111513E+16

...

65640 273 12:00:00 1.794810E-02 1.987897E+04 1.987899E+04 1.111513E+16
WRT_HIS - wrote history fields (Index=1,1) into time record = 0000548
65641 273 12:06:00 1.794841E-02 1.987897E+04 1.987899E+04 1.111513E+16

[jcwarner]

is the time variable in the netcdf file correct? Make sure the time is continuously increasing, no backward steps.

[fragoso]

Yes. There are some backward steps on the sms_time variable. I´ll correct that.

Thank you!

Mauricio Fragoso
PROOCEANO - Brasil