mpas_timekeeping.F: Fortran runtime error - Bad integer for item 1 in lnput

[makinde]

Hi,

please I am a newbie to MPAS and model simulation, so please forgive any of my mistakes or omission or anything I could be violating. Thanks
I have been trying to run a variable resolution model for a year starting at 2009-01-01_00 through to 2009-12-31_00 at 6hrs interval. I have downloaded necessary meteorological files, run the static initialization successfully without any error, done the initial condition without any error. In my mpas_init I have files x6.999426.grid.nc, x6.999426.static.nc, x6.999426.init.nc, and x6.999426.sfc_update.nc and I have well linked all these files into the mpas_atmos. I have set up the namelist.atmosphere and the streams.atmosphere.
Now I try to run the atmospher_model, it ran for a few seconds and terminated without any error in the log.atmosphere.0000.out but the atmos_stderr file contains the following error:

Error termination. Backtrace:
At line 1495 of file mpas_timekeeping.F
Fortran runtime error: Bad integer for item 1 in list input

I should have attached the error file and log file but I don't know how to do that on this platform.
please what could be wrong?


Thanks

 

[mcurry]

It seems like it might be an syntax error in specifying the time in the namelist.atmosphere. Posting the it (namelist.atmosphere) as well as the log.atmosphere.0000.out would be helpful.

To attach a file on this forum click 'Post Reply' or 'Full Editor & Preview' below this post. Then, below the text edit box you should see a a box with tabs for 'Options' and 'Attachments'. Click 'Attachments' and then 'Add files' to add the above files.

 

[makinde]

Thanks mcurry,
please find in the attach the log file, the namelist and the stream files.

Thanks

 

[mcurry]

Thanks for attaching those files. I'm not sure if this is the cause of the problem, but I am noticing that your horizontal length scale (config_len_disp in the nhyd_model section of the namelist.atmosphere) appears to be too large for the 60-10km mesh as well, your time step (config_dt) might also be a little too large.

The horizontal length scale, config_len_disp, should be set to the smallest grid distance in a mesh in meters. In your case this should be 10000. Likewise, the smallest grid distance of a variable revolution mesh should determine the time step. A good rule of thumb for finding an inital timestep is to multiply the smallest grid distance by 6. So in your case config_dt should be 60.

If there was a reason that you choose the values that are in your namelist please let me know. Otherwise see if the error occurs when running with the above values.

I'd also double check that the xtime field in x6.999426.init.nc contain the same time as in config_start_time. Other then config_len_disp and config_dt, I'm not seeing anything else out of the ordinary.

 

[makinde]

Thanks so much Mcurry for the great help, I did as you pointed out change those values and run it again, it ran further than before but stop again with the same error in the atmos_stderr.
While I was bothered and concerned, a friend of mine going through the streams.atmosphere discovered that I was using surface update for my SST but set the input_interval in stream to none, so upon changing it to 21600, as the time step of the SST data is 6hourly and keeping all the values as you pointed out above, it ran.
Thank you so much, I really appreciate all your help and contributions, am so happy to have it running being my first experience with MPAS simulation.
Thanks.

 

[mcurry]

Yes! Enabling SST updates must be done in two places. This often is missed as it was by me. Glad you got it working!

 

[mgduda]

There was a separate, follow-on question regarding simulation rate that I've moved to a new thread.