Opened 12 years ago

Closed 12 years ago

#122 closed task (fixed)

start dumps request

Reported by: gnp Owned by: willie
Component: Data Keywords: start dumps
Cc: g.n.petersen@… Platform:
UM Version:

Description

Dear all,

The GFDex group at UEA is interested in doing simulations for a case in March this year.
We're not in a hurry to start the simulations but I want to make sure we have the start dumps that we need when the time comes.

Could you please retrieve and store on hpcx the global start dumps for the following times:

09.03.2008 12UTC (and if available 18UTC)
10.03.2008 00UTC (and if available 06UTC)
10.03.2008 12UTC (and if available 18UTC)

both T+0 and T+3

Thanks,

Nina


Gudrun Nina Petersen
School of Environmental Sciences
UEA

Change History (15)

comment:1 Changed 12 years ago by gnp

Could you also please save NAE start dumps for the same dates?

Thanks,

Nina


Gudrun Nina Petersen
School of Environmental Sciences
UEA

comment:2 Changed 12 years ago by lois

Hello Nina, William McGinty? now has an account at the Met Office to take over the role of Chang Gui Wang in getting start data. He is still learning to ropes so I was glad you said you were not in too much of a hurry as you will be the guinea pigs in setting up this service.

Hopefully you may not really need the NAE start files for the periods of the global start data as you should be able to reconfigure your global start data using some NAE ancillary files that are either on HPCx already or available via the NCAS UM ancillary files service run at the BADC. Helen Dacre is currently running the NAE like this as she is starting from a start date unavailable via the Met Office and so she is using an ECMWF start dump. However Helen has had problems, which seem to have nearly been resolved, and so we need to understand why there are problems running the NAE from a global reconfigured dump with ancillary files when we can run the NAE with no problems from a UKMO NAE start file.

Lois

comment:3 Changed 12 years ago by willie

  • Owner changed from um_support to willie
  • Status changed from new to assigned

comment:4 Changed 12 years ago by willie

  • Summary changed from Start dumps request to Work started

Nina,

I am now looking at this full time, having just got my Met Office MASS account set up. Apologies for the delay.

Regards,
Willie

comment:5 Changed 12 years ago by gnp

Willie,

no problem, we're still not in a hurry.

Cheers,

Nina

comment:6 Changed 12 years ago by willie

  • Summary changed from Work started to start dumps request

comment:7 Changed 12 years ago by willie

Nina,
I have transferred some start dumps (global) into the directory Met_O_MASS in the dumps/weather directory on HPCx. I have not fully checked this out (I have done checksum and looked at it with xconv and it seems OK). The transfer of data from the Met Office computer to HPCx took in excess of 43 hours. I am not proposing to provide NAE data unless you insist, due to the timescales involved as others are requesting data. Let me know if you really need the NAE data.

If you could also let me know if the data has proved correct and useful.

regards
Willie

comment:8 Changed 12 years ago by gnp

Hi Willie,

I had a look in both
/hpcx/work/n02/n02-ncas/umx/dumps/weather/global
and
/hpcx/devt/n02/n02-ncas/umx/dumps/weather/global

but couldn't see any 2008 start dumps nor a Met_O_MASS directory.

Is there a third place I should be looking?

All our current start dumps are in
/hpcx/work/n02/n02-ncas/umx/dumps/weather/global
It would be preferable if the new ones could be stored there also.

When I've found the start dumps, I'll do a short global run with one of
them and assuming that one is OK this is probably all we need.

Cheers,

Nina

comment:9 Changed 12 years ago by willie

Nina,
I have been putting the start dumps on /hpcx/home/n02/n02/umx/dumps/weather (NB ../home/.. as this is backed up) in the directory Met_O_MASS. The work directory is not backed up.

Sorry for the confusion.

Willie

comment:10 Changed 12 years ago by gnp

Hi Willie,

I found them.
Just out of curiosity, do you know what's the difference between the 2007 dumps and the 2008?
There is a clear size difference:
-rw-r-xr-x 1 umx n02-ncas 1243332608 Apr 04 2007 2007022812.T+0
-rwxr—r— 1 umx n02 1555169280 Apr 21 13:45 20080310_qwqg12.T+0

I'll to a trial run soonish.

Thanks,

Nina

comment:11 Changed 12 years ago by willie

Hmmm, I see what you mean about the sizes. I am investigating …
Willie

comment:12 Changed 12 years ago by willie

Hi Nina,
OK. My method for getting the data agrees with the previous method. Comparing the two files with xconv shows that they both have the same fields and sizes except for the very last field:
the 2007 file has it as 640x481x1 NET DN SW RAD FLUX: OPEN SEA: SEA MEAN
the 2008 file has it as 640x481x50 Stash code=321

This accounts for the size difference.

Let me know how your trial run works out.
Regards
Willie

comment:13 Changed 12 years ago by willie

Nina,

Stephen Outten has had some problems with the start dump data, due to peculiar fields. I have given Stephen details of a solution involving reconfiguration and some user STASH masters. Please see Stephen for details, or contact me.

Regards
Willie

comment:14 Changed 12 years ago by willie

Nina,

I have now updated all the start dumps in the dumps/weather/Met_O_MASS/global directory as of noon on the 28th May. This was to overcome problems discovered by Stephen Outten.

Regards
Willie

comment:15 Changed 12 years ago by willie

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