Hello, when I run the CCTM module of CMAQv5.0.2, I encounter the error message “netCDF error number -40 Starting time not on file”. I have checked the dates in MCIP and OCEAN files, and the dates are correct. What could be the possible reason for this? Thanks for you help!
Message says it. Look at the log: ACONC_1 starts June 30, 2018, with NO TIME STEPS (maximum record number 0)
Please post the full log file and run script. The portion of the log file you posted suggests that you may be trying to restart the model for August 2, 2018 since that the is the time stamp in the CGRID file, and yet the names of the ACONC and CGRID files imply a simulation date of June 30, 2018. Or are you trying to simulate June 30, 2018 but are providing an August 2, 2018 restart file?
Thank you very much for your response. However, as a new user, I am unable to upload attachments. So, I will reply to you with a picture of the CCTM run script. Also, could you please tell me where the restart file is specified? I haven’t noticed this file before.
Hello, I can now upload attachments. Here are the log file and the script I used to run CCTM, due to the large size of the file, I split the log files into two parts… I noticed that the dates in the CGRID file do not match the dates I intended to run. So I changed to a new output path, but I still encountered the error message “netCDF error number -40.” Could this be a problem with the settings?
d01_2.txt (5.3 MB)
d01_1.txt (7.4 MB)
cctm.txt (14.5 KB)
Thanks for sharing these files.
The error causing this problem very likely is at the very end of your run script:
time mpirun -np $NPROCS $BLD/$EXEC
time $BLD/$EXEC
The second line (“time $BLD/$EXEC”) shouldn’t be there, otherwise you are telling the script to run CCTM twice, once with mpi and once without.
Your run script sets up a 33 day CCTM simulation starting June 30, 2018, and your two log files show that the run progresses through all of these days and successfully finishes at the end of August 1, 2018. You can look at lines 131370 - 131382 of d01_2.txt to confirm this:
Timestep written to S_CGRID for date and time 2018214:000000The elapsed time for this job was 068096 seconds.
--->> Normal Completion of program DRIVER on PE 000 <<--- Date and time 0:00:00 Aug. 2, 2018 (2018214:000000) >>----> Program completed successfully <----<<3257220.198u 67849.375s 18:53:53.32 4887.4% 0+0k 0+122913128io 0pf+0w
After that, the run script tries to start another CCTM run with the erroneous " time $BLD/$EXEC" command, and that run then fails because the output files already exist and are opened as read-only so CCTM cannot write to them and the script then exits after the fail of that second run attempt.
I admit that I don’t quite understand why the existing A_CONC_1 file being opened as read-only has no existing time steps while the existing CTM_VIS_1 file is reported to have the expected 792 time steps from the successful first run and the S_CGRID shows the expected August 2, 2018 00:00 end time of the first run. But in any case, I would recommend to clean up your output directory, remove the line “time $BLD/$EXEC” from your run script, and try again since the bulk of your two log files show that you can successfully run the entire June 30 - August 1, 2018 time period.
2 Likes
Thank you very much for your patient response. The issue has been resolved.
1 Like