Opened 3 weeks ago

Closed 10 days ago

#2839 closed help (fixed)

dump file for UM8.2 for 1st June needed

Reported by: jcrook Owned by: willie
Component: UM Model Keywords: fix, UM, file, mule
Cc: Platform: ARCHER
UM Version: 8.4

Description

Hi Willie

Some time ago we discovered a problem with our original vegetation fraction ancillary file - it had 2 grid cells over land with no vegetation fractions set. This caused the surface temperature to be set to zero for these grid cells. Although your simulation xmrjc ran ok for 122 days we did not spot this problem until we tried to use the dump files to reconfigure other simulations. I think I changed the starting dump file to be the file you used to start the 4km simulation xmrjc (ie xmpbc.astart. I also changed the way some ancillaries were configured. This was documented in ticket #2524.

I have been able to run lots of simulations from 5th April 2014 by running one simulation from 1st April using the xmpbc.astart file and then using that simulation's dump file from 5th April.

Now I want to run some simulations from 1st June 2014 but I don't think I can use the xmrjc dump file for 1st June because it has these 2 grid cells with zero surface temperature. Ideally I need a way to set these 2 grid cells surface temperature to something other than zero (eg the mean of the surrounding grid cells) in the dump file for 1st June for xmrjc. Is there a way to do this? If not what dump file can I use? I don't want to run a simulation from April through to 1st June at 4km as it will take too long and I don't have the AUs on ARCHER to do this.

Regards
Julia

Attachments (2)

vera.py (1.7 KB) - added by willie 3 weeks ago.
vera.py.jpg (302.7 KB) - added by willie 2 weeks ago.

Download all attachments as: .zip

Change History (12)

comment:1 Changed 3 weeks ago by willie

Hi Julia,

What's the path to the xmrjc dump for 1st June - I need to remind myself about this.

Willie

comment:2 Changed 3 weeks ago by jcrook

Its at /home/n02/n02/jcrook2/work/VERA/um/xmrjc/

I was thinking may be I could use one of my April dump files and change its date to 26th May then run for 5 days to get a 1st June dump file which I could then use for my ensemble. I don't know how quickly everything would adjust.

Julia

comment:3 Changed 3 weeks ago by willie

Hi Julia,

Do you know the lat/long of the zero deg points?

Are the surrounding values minimums or maximums or somewhere in the middle?

It may be possible for me to write a bit of python that fixes it.

Willie

comment:4 Changed 3 weeks ago by jcrook

Hi Willie

The problem happens at data[176,314] and data[177,314]. I am a bit wary of using the xmrjc dump files at all because I don't know if the Ts variable is the only one affected. The 12km simulation did not have the problem so can we reconfigure a 4km run from a 12km dump file?

comment:5 Changed 3 weeks ago by willie

Hi Julia,

OK. Reconfiguring from 12km to 4km will work but will introduce some noise in all the fields compared to using a good 4km dump.

Willie

Changed 3 weeks ago by willie

comment:6 Changed 3 weeks ago by willie

  • Cc fix, UM, file, mule added

Hi Julia,

I have now corrected the dump file using the attached python code. The modified dump is at /work/n02/n02/wmcginty/xmrjca.da20140601_00.fixed. I just inserted 297.3 deg K at the two points in question use mule operators.

Willie

comment:7 Changed 2 weeks ago by jcrook

Thanks Willie

I have started a run using this fixed dump file and it has done the reconfiguration and is now running so I guess it is ok. I also tried reconfiguring another run using our equivalent 12km dump file but it didn't like it because the domain is ever so slightly smaller for the 12km simulation and it says the target domain is outside the source domain.

Changed 2 weeks ago by willie

comment:8 Changed 2 weeks ago by willie

Added a before and after picture for completeness.

comment:9 Changed 10 days ago by willie

  • Cc fix, UM, file, mule removed
  • Keywords fix, UM, file, mule added

comment:10 Changed 10 days ago by willie

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