Opened 4 years ago

Closed 4 years ago

#2199 closed help (answered)

Tropospheric+Isoprene Chemistry in vn8.4 Release Job

Reported by: webber24 Owned by: um_support
Component: UKCA Keywords: UKCA, Chemistry, Vn8.4
Cc: Platform: ARCHER
UM Version: 8.4


Dear CMS,

I have a vn8.4 job (xngxx) which I had run previously with Stratospheric+Tropospheric Chemistry. I attempted to build and reconfigure the same job with the following radio button ticked:

Atmosphere >

Scientific Parameters and Sections >

Section by section choices >

UKCA Chemistry and Aerosols >

Tropospheric+Isoprene Chemistry

However I received the following error in my .leave file (/home/n02/n02/webber24/output/xngxx000.xngxx.d17163.t201511.leave):


???!!!???!!!???!!!???!!!???!!!???!!! ERROR ???!!!???!!!???!!!???!!!???!!!???!!!?
? Error in routine: UKCA_CHEM1_INIT.check_dims
? Error Code: 91
? Error Message: chch_defs: size not equal to jpspec
? Error generated from processor: 0
? This run generated 10 warnings

My understanding of this error is that the number of chemical species being defined in ukca_chem_tropisop.F90 is 64, as opposed to 67 being previously defined in the strattrop version. My question is how to go about setting 64 universally and why this has not been altered by the radio button switch?



Change History (6)

comment:1 Changed 4 years ago by luke

Hi Chris,

Unfortunately, the TropIsop (AKA CheT) scheme you are trying to use is not supported at vn8.4. Their are numerous changes that need to be made between the two schemes. Also, the rates could be slightly different, as only the StratTrop rates may have been updated.

Some information can be found here, although the jobs will be quite old now and may not have been set-up to run on ARCHER/Monsoon2:

I would suggest picking a TropIsop/CheT job and compare the settings to see what to take. I suspect that the issue you are seeing is to do with using hand-edits to turn tracers on, but possibly it's something else.

Can I ask why you want to use this scheme over StratTrop/CheST?


comment:2 Changed 4 years ago by luke

Also, I should add that in the past I found running TropIsop for more than a few years wasn't possible, as the model became unstable. This is partly why it is not supported.


Last edited 4 years ago by luke (previous) (diff)

comment:3 Changed 4 years ago by webber24

Hi Luke,

Thanks for that, the scheme is preferred as it was used in a previous job, with which we are trying to compare. I think it's back to the drawing board, also thank you for the advice re. stability.


comment:4 Changed 4 years ago by luke

What version was the previous job at?

In terms of stability, it is always possible that the newer compilers/architecture might be OK, however, as I never tracked down exactly what was going on, I don't want to say it wouldn't work for certain.


comment:5 Changed 4 years ago by webber24

Hi Luke,

The previous job was run at vn 8.2 on Monsoon. It originated from Mohit's xhmag job, so if there is an 8.2 job with Tropisop/CheT on ARCHER then that would be perfect.


comment:6 Changed 4 years ago by luke

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

Hi Chris,

Sorry for the delay in my reply.

I'm afraid that I can't recommend any vn8.2 jobs - all supported jobs are at vn8.4, and use StratTrop/CheST, rather than TropIsop/CheT. There was some development work done, but this was discontinued due to the problems I've already mentioned and also the fact that we wanted to support fewer versions.

You could try getting TropIsop/CheT to work, but you'd need to do this yourself, and if you run into problems my response might be "please use CheST" I'm afraid. You should look at these jobs and see if you can modify something to match your needs, but please be aware of the limitations I've discussed above.

Sorry about this. I'll close this ticket now. Please open another if you have any further issues.


Note: See TracTickets for help on using tickets.