Opened 2 years ago

Closed 2 years ago

#2215 closed help (answered)

Global model running for 5 days

Reported by: sam89 Owned by: um_support
Component: UM Model Keywords:
Cc: Platform: Monsoon2
UM Version: 8.2

Description

Hi

I have two jobs xnjjh and xnjjp. They are both the same and I can't see anything when looking at the differences between the two as to why this issue is happening but the T+ X times do not match between the two jobs. As far as I can see I have set them both running from 18 hrs with an 18 hrs start dump but the output from the xnjjh job still seems to give times as if it is starting from 00 hrs. I was wondering if you could work out why this is happening? I checked my reconfig jobs and they both also start from 18 UTC so I am none the wiser.

Change History (11)

comment:1 Changed 2 years ago by sam89

I checked the output files and the xnjjh job seems to have 18 hrs as 0.75 days into the run whereas xnjjp has it as 0 days into the run. So something is not right between the two but I cannot work it out.

Last edited 2 years ago by sam89 (previous) (diff)

comment:2 Changed 2 years ago by willie

Hi Sam,

Both jobs xnjjh (w/o IAU) and xnjjp (with IAU) have run successfully as far as I can see. What files are you looking at?

Regards,
Willie

comment:3 Changed 2 years ago by sam89

Hi Willie

Sorry about that I meant to say they've both ran but the xnjjh job in the output files seems that it is still thinking the run is starting from 00 but the xnjjp is starting at 18. The output says for instance that time 06 UTC on 6th July is T+30 for the xnjjh job but for the xnjjp job it says T+12.
I can't work out why they both don't say T+12 since they both use the same start dump and are started from 18 UTC on 5th july.

Thanks

Sam

comment:4 Changed 2 years ago by sam89

So for instance see xnjjh.pp0

comment:5 Changed 2 years ago by willie

Hi Sam,

Yes, the pp0 file for h has a reference date of 2012/07/05:00:00 while p has 2012/07/05:18:00. I'm not sure but I think this may be due to the IAU scheme. You could test this by making a copy of p and running it with the IAU off and then compare with h - do you still get the reference date difference?

Regards
Willie

comment:6 Changed 2 years ago by sam89

Hi Willie

Thanks I will give that a check. Would it make a difference to the output though?

Thanks

Sam

comment:7 Changed 2 years ago by sam89

I also keep forgetting to ask. Could you check xnjjp000.xnjjp.d17181.t045701.leave for the IAU file the output has * for some of the height levels instead of numbers and I was wondering if the IAU file had been added properly to the run?

Thanks

Sam

comment:8 Changed 2 years ago by sam89

Hi Willie

Don't worry I asked Sue and she said it seems to have run fine.
Thanks

Sam

comment:9 Changed 2 years ago by sam89

Hi Willie

I reran job xnjjp without the iau switched on and that is job xnjjg. It seems that both of them are starting at timestep 145 though when looking at the output files. I have no idea why this would be though as the start dump starts at 18 utc which should mean the run starts at this time too so 18 utc would be timestep 0 right? I have definitely set the job up to start at 18 too and not 0 so I can't work out why it wouldn't have 18:10 as timestep number 1. Do you have any idea why this would be as I am completely puzzled!

Thanks

Sam

comment:10 Changed 2 years ago by willie

Hi Sam,

This indicates that switching on the IAU scheme causes the change in reference date. Whether this is an error or a matter of interpretation is not clear to me. You should review the way you set up the IAU scheme. See UMDP 31: Incremental Analysis Update (IAU) Scheme for more information.

Regards
Willie

comment:11 Changed 2 years ago by willie

  • Resolution set to answered
  • Status changed from new to closed
Note: See TracTickets for help on using tickets.