#2968 closed help (answered)

Global JULES tutorial

Reported by: NoelClancy Owned by: pmcguire
Component: JULES Keywords:
Cc: Platform:
UM Version:


When restarting a global run at step 13

"You would need to change the dump file name in jules_initial to correspond to the last dump file in your output_dir."
The last dump file in my output_dir is:

Therefore what sould the file path in jules initial be?

Should it be:

Change History (8)

comment:1 Changed 15 months ago by pmcguire

  • Cc mcguirepatr@… removed

comment:2 Changed 15 months ago by pmcguire

  • Status changed from new to accepted

comment:3 Changed 15 months ago by pmcguire

Hi Noel:
The filename with the path for the UK-only run should be:

Why do you want to restart this run? This is a UK-only run, not a global run, so it was able to finish without needing a restart.
The last year of the driving data is 2012 (the first year is 1979).
Can you use ncview to study the output NETCDF files for the UK only run?

it looks like your global run didn't finish yet.
So if you would choose the filename with the path for the global run as something like:

Does this help?

comment:4 Changed 15 months ago by NoelClancy

ok thanks,
In the global run (u-bl019) what should I do in section [namelist:jules_spinup]?

Because there are dump files for all 35 spin up cycles, even after the run initiated in Edinburgh, I assume I comment out line 640 and uncomment line 644 as follows:



I've changed the file name to what you recommend and changed the max_spinup_cycles to zero

but I'm getting the following error

ERROR: file not found: /home/users/nmc/cylc-run/u-bl019/log/job/1/fcm_make/NN/job.err
Traceback (most recent call last):

File "/apps/contrib/metomi/cylc-7.8.1/bin/cylc-cat-log", line 439, in <module>


File "/apps/contrib/metomi/cylc-7.8.1/bin/cylc-cat-log", line 435, in main

tmpfile_edit(out, options.geditor)

File "/apps/contrib/metomi/cylc-7.8.1/bin/cylc-cat-log", line 265, in tmpfile_edit

modtime1 = os.stat(tmpfile).st_mtime

TypeError?: coercing to Unicode: need string or buffer, int found

comment:5 Changed 15 months ago by NoelClancy

Can you use ncview to study the output NETCDF files for the UK only run?
Yes, I can but for some variables it's blank, the variables it shows output for the map of the UK does not look very refined.

Maybe I should "cdo mergetime" the files and then "ncview" the merged time series?

comment:6 Changed 15 months ago by pmcguire

Hi Noel:
Which variables are blank for the UK JULES runs?
What do you mean by 'not look[ing] very refined'?
Maybe you can change the color range plotted with the range button in ncview?
There are other things you can do too with ncview.
Does changing the color range help? Or are some of the vsriables still missing?
Are the variables missing/blank for all the time steps?

comment:7 Changed 15 months ago by pmcguire

Hi Noel
For these GL6R-style global suites, the location of the log files (job.err and job.out) are not in their normal location, so the Cylc GUI can't find them. The new location of the log files (for these GL6R-style global suites) is defined in your suite.rc file and it is located on the scratch drive. What error messages do you see in your log files?

It looks like you changed the spin_up namelist properly. But you also need to change the jules_time namelist since you're not restarting in 1979.

I highly recommend that people should rename their log files if they are doing a restart. That way you will have different log files for each restart.

Does that help?

comment:8 Changed 13 months ago by pmcguire

  • Resolution set to answered
  • Status changed from accepted to closed

I am closing this ticket.

Note: See TracTickets for help on using tickets.