Computing problem!

[langlangxie]

Hi,I have a problem in computing ROMS.
This is :
WRT_AVG - error while writing variable: zeta
into averages NetCDF file for time record: 1

Elapsed CPU time (seconds):

Node # 0 CPU: 1009.411

Node # 4 CPU: 1009.387 ROMS/TOMS - Output NetCDF summary for Grid 01:
Node # 5 CPU: 1009.386 Node # 10 CPU: 1009.406 Node # 3 CPU: 1009.394 Node # 1 CPU: 1009.411 Node # 8 CPU: 1009.410 Node # 12 CPU: 1009.410 Node # 9 CPU: 1009.402 Node # 2 CPU: 1009.394 Node # 11 CPU: 1009.405 Node # 14 CPU: 1009.402 Node # 7 CPU: 1009.387 Node # 6 CPU: 1009.387 Node # 13 CPU: 1009.409 Node # 15 CPU: 1009.400
number of time records written in AVERAGE file = 00000001















Analytical header files used:

ROMS/Functionals/ana_btflux.h

ROMS/TOMS - Output error ............ exit_flag: 3


ERROR: Abnormal termination: NetCDF OUTPUT.
REASON: NetCDF: Numeric conversion not representable

Can you help me ?

[kate]

Have you set NINFO to 1? If you do, you will get a diagnostics report each timestep. If the model is blowing up, you would then see it there first. If NINFO is large, you can get a NaN and ROMS will keep running. I believe you have NaN or some such number in your zeta field, which is why it won't write to the NetCDF file.

[langlangxie]

Have you set NINFO to 1? If you do, you will get a diagnostics report each timestep. If the model is blowing up, you would then see it there first. If NINFO is large, you can get a NaN and ROMS will keep running. I believe you have NaN or some such number in your zeta field, which is why it won't write to the NetCDF file.

hi,kate.
My NINFO is 51840,and I want to understand how to make the NINFO,can you give some your understanding.
Thank you very much!

[kate]

NINFO is set in your ocean_xx.in file. Try setting it to 1. If you can get things to run stably, then you can set it to a large value again if you like. I always leave it at 1 and put up with the extra output.

[langlangxie]

NINFO is set in your ocean_xx.in file. Try setting it to 1. If you can get things to run stably, then you can set it to a large value again if you like. I always leave it at 1 and put up with the extra output.

Thanks a lot. Thanks a million!

[langlangxie]

Have you set NINFO to 1? If you do, you will get a diagnostics report each timestep. If the model is blowing up, you would then see it there first. If NINFO is large, you can get a NaN and ROMS will keep running. I believe you have NaN or some such number in your zeta field, which is why it won't write to the NetCDF file.

Hi kate
Can you tell me how to find zeta? Is there zeta in clim.nc or ini.nc? Then how can i get the value of zeta?
Thank you!

[kate]

Zeta is the free surface elevation and it evolves during a model run. It would be in your ini file and also in the restart file, plus whatever other outputs you've asked for. I suggest ncview to look at the netcdf input and output files.

[langlangxie]

Zeta is the free surface elevation and it evolves during a model run. It would be in your ini file and also in the restart file, plus whatever other outputs you've asked for. I suggest ncview to look at the netcdf input and output files.

Thanks for your help! Firstly ,I got a error is "WRT_AVG - error while writing variable: zeta into averages NetCDF file for time record: 1",this time my NINFO=51840. Then I make the NINFO to 5,the error is " WRT_AVG - error while writing variable: zeta into averages NetCDF file for time record: 5",why is the error happening.What is the best value of NINFO?
And I ncview my netcdf input and output files,there is zeta and ssh.I think my zeta is realy.

[kate]

The best value for NINFO is 1 (as above). That way, ROMS will check for NaN and other such values every time-step. Your model is going bad and you need to figure out where and why.

[langlangxie]

The best value for NINFO is 1 (as above). That way, ROMS will check for NaN and other such values every time-step. Your model is going bad and you need to figure out where and why.

Thanks!
I am finding my errors in my ROMS.can you give me some directions of going bad?
Today,I get a result ,and I get 20 avg files,likes "creating average file, Grid 01: /public/userspace/langlangxie/south/clim_avg_M_d8_0020.nc",why it is 20?

[kate]

Are you changing NAVG as well as NINFO? You should read up on these parameters - they are described briefly at the bottom of the ocean.in file.

As for ROMS blowing up, there are a few tips on the wiki.

[langlangxie]

Are you changing NAVG as well as NINFO? You should read up on these parameters - they are described briefly at the bottom of the ocean.in file.

As for ROMS blowing up, there are a few tips on the wiki.

Dear kate
Firstly,thanks again!
My NAVG is not equal to NINFO.NAVG=4320,and NINFO=51840.
When I make as follows: NTIMES == 51840 ！1 years
DT == 60.0d0
NDTFAST == 600
I get a error: LBC_GETATT - error while reading global attribute: NLM_LBC
in restart file: scs_ini.nc
call from: get_state.F
Probably global attribute was not found ...
restart file needs to be generated by ROMS version 3.6 or higher
Alternatively, you may use NO_LBC_ATT at your own risk!
When DT=600,ROMS can running stable,but I make DT=60,there is error above.
Can you help me solve my problem!
Thanks!

[kate]

I told you to try NINFO=1.

With dt=60 and that many steps, you are asking it to run for 36 days, saving an averages file every 3 days. You shouldn't get 20 averages files, just 12. Right?

I've never heard of anyone using ndtfast as large as 600.

You must have told it that your initial file is a restart file. What is your NRREC?

[langlangxie]

I told you to try NINFO=1.

With dt=60 and that many steps, you are asking it to run for 36 days, saving an averages file every 3 days. You shouldn't get 20 averages files, just 12. Right?

I've never heard of anyone using ndtfast as large as 600.

You must have told it that your initial file is a restart file. What is your NRREC?

I make NINFO=1,it can running stable.And I get 12 averages.
Thanks!