Welcome, Guest
Username: Password: Remember me
Forum header
  • Page:
  • 1
  • 2

TOPIC: Running a HIRLAM model with internal nesting

Running a HIRLAM model with internal nesting 8 years 2 months ago #420

  • Xiaohua Yang
  • Xiaohua Yang's Avatar
  • OFFLINE
  • Administrator
  • Posts: 195
  • Thank you received: 4
Suleiman asked:

> Hello!
> I want to run hirlam with 3 nested domain
> the first domain i ran and now in which parameters in which scripts should i
> change to use main doamin as boundry ???
>

As far as I recall one needs to do following:
1. edit scripts/Env_expdesc to declare
BDSTRATEGY=same_cycle
BDINT=01
(the above assumes that when you run your outer domain, you have made changes in Env_domain to specify an hourly history output in WRITUPTIMES,
else you need to specify BDINT here according to whatever data interval you can provide from outer domain)
2. edit Env_input to specify
NESTING=yes
and to specify LBC data source and path, such as
INBDEXP (name of outer model experiment)
INBDDIR (data path of the outermodel, such as
$HL_ARC/$INBDEXP
where HL_ARC is normally your experiment result archive directory

Other then these there are already settings for NESTING=yes scenario in scripts like Bdries, Boundary, MakeStrategy, makeStrategy.pl. Check about these if in doubt.

Re: Running a HIRLAM model with internal nesting 8 years 2 months ago #421

Thanks for your answer
and what about analysis - must I turn the ANALYSIS to NONE or i can choose any analysis scheme ?

Re: Running a HIRLAM model with internal nesting 8 years 2 months ago #422

  • Xiaohua Yang
  • Xiaohua Yang's Avatar
  • OFFLINE
  • Administrator
  • Posts: 195
  • Thank you received: 4
Suleiman Mohammad Mostamandy wrote:
Thanks for your answer
and what about analysis - must I turn the ANALYSIS to NONE or i can choose any analysis scheme ?

no, one does not have to put ANALYSIS=NONE, and in fact "NONE" shall normally not be considered with the current HIRLAM forecast system. NONE in HIRLAM means that only downscaling from upper air boundary is done, for the surface, it will cylce from surface climate. I think one can think instead of two alternatives:
1. (In Env_expdesc) you choose ANALYSIS=NOUA,
which will let you use upper air fields from nesting model, but it will still carry oout surface analysis.
2. You choose ANALYSIS=3DVAR which will then do also its own upper air analysis.

Most likely you are using nesting strategy for small domain, for which only limited number of observations are available, so the above option 1) maybe better to use.

Re: Running a HIRLAM model with internal nesting 8 years 2 months ago #423

Thanks again!
now i faced with new problem - after editing the scripts i have got this message:

Array @list missing the @ in argument 1 of push() at /home/hirlam/scratch/hl_home/exp_rshu/lib/scripts/util.plib line 84.
Array @list missing the @ in argument 1 of push() at /home/hirlam/scratch/hl_home/exp_rshu/lib/scripts/util.plib line 90.

Re: Running a HIRLAM model with internal nesting 8 years 2 months ago #446

I still have problem with running hirlma in nest mode !
so the problem with LBCn (1..10)
i attached with this message the log files with Exp_input

thanks in advance for your help !!!

File Attachment:

File Name: exp_rshu2.zip
File Size: 41981
Last Edit: 8 years 2 months ago by Suleiman Mohammad Mostamandy.

Re:Running a HIRLAM model with internal nesting 8 years 2 months ago #447

  • Eoin Whelan
  • Eoin Whelan's Avatar
  • OFFLINE
  • Gold Boarder
  • Posts: 195
  • Thank you received: 34
Suleiman

I have carried out a quick test on my PC using a nesting strategy using hirlam-7.4beta1:

I think the output you are getting from util.plib is significant. (I do not get this with my tests). Perhaps you could add some diagnostic warn statemnets in this file.

Is it possible you are trying to read an empty file?

The output in your Boundaries.1 log files conatain:
.
.
.
UNABLE TO FIND FIELD LN(PS) 152
UNABLE TO FIND FIELD/LEVTYP/LEV 1 105 0
.
.
.
My experience in the past is that you get this error if the input file is corrupt/empty. I think surface pressure is the first field to be read in.

Eoin

Re:Running a HIRLAM model with internal nesting 8 years 2 months ago #448

Thanks for your answer :
could you please tell which files needed as boundary ???
i did wgrib fc*md and couldn't find field with id 152. but in fc* there are fileds with 152 id kpds5=152:kpds6=109

here is the list of outer domain files :
as you can see all files have normal size.

total 514808
-rw-r
. 1 hirlam hirlam 4674240 Nov 7 12:04 ma201011050000.mars
-rw-rw-r--. 1 hirlam hirlam 4728000 Nov 7 12:04 ma201011050000m
-rw-rw-r--. 1 hirlam hirlam 3360000 Nov 7 12:04 ma201011050000f
-rw-rw-r--. 1 hirlam hirlam 192000 Nov 7 13:34 ob2010110500f1
-rw-rw-r--. 1 hirlam hirlam 84000 Nov 7 13:34 ob2010110500f5
-rw-rw-r--. 1 hirlam hirlam 1092000 Nov 7 13:34 an20101105_00+000ve
-rw-rw-r--. 1 hirlam hirlam 1116000 Nov 7 13:34 fc20101105_00+000ve
-rw-rw-r--. 1 hirlam hirlam 228000 Nov 7 13:34 fc20101105_00+003ve
-rw-rw-r--. 1 hirlam hirlam 1128000 Nov 7 13:34 fc20101105_00+006ve
-rw-rw-r--. 1 hirlam hirlam 228000 Nov 7 13:34 fc20101105_00+009ve
-rw-rw-r--. 1 hirlam hirlam 1128000 Nov 7 13:34 fc20101105_00+012ve
-rw-rw-r--. 1 hirlam hirlam 228000 Nov 7 13:34 fc20101105_00+015ve
-rw-rw-r--. 1 hirlam hirlam 1140000 Nov 7 13:34 fc20101105_00+018ve
-rw-rw-r--. 1 hirlam hirlam 240000 Nov 7 13:34 fc20101105_00+021ve
-rw-rw-r--. 1 hirlam hirlam 1140000 Nov 7 13:34 fc20101105_00+024ve
-rw-rw-r--. 1 hirlam hirlam 4553763 Nov 7 13:34 OBST20101105_00.tar.gz
-rw-rw-r--. 1 hirlam hirlam 38441 Nov 7 13:34 OF20101105_00.tar.gz
-rw-rw-r--. 1 hirlam hirlam 43000 Nov 7 13:34 ve20101105_00
-rw-rw-r--. 1 hirlam hirlam 102035928 Nov 7 13:34 ACMA20101105_00+000
-rw-rw-r--. 1 hirlam hirlam 8854012 Nov 7 13:34 ACMA20101105_00+000.gz
-rw-rw-r--. 1 hirlam hirlam 2928000 Nov 7 13:34 an20101105_00+000
-rw-rw-r--. 1 hirlam hirlam 852000 Nov 7 13:34 an20101105_00+000md
-rw-rw-r--. 1 hirlam hirlam 157 Nov 7 13:34 bdsy
-rw-rw-r--. 1 hirlam hirlam 14136000 Nov 7 13:34 fc20101105_00+000
-rw-rw-r--. 1 hirlam hirlam 924000 Nov 7 13:34 fc20101105_00+000md
-rw-rw-r--. 1 hirlam hirlam 14148000 Nov 7 13:34 fc20101105_00+001
-rw-rw-r--. 1 hirlam hirlam 948000 Nov 7 13:34 fc20101105_00+001md
-rw-rw-r--. 1 hirlam hirlam 14148000 Nov 7 13:34 fc20101105_00+002
-rw-rw-r--. 1 hirlam hirlam 948000 Nov 7 13:34 fc20101105_00+002md
-rw-rw-r--. 1 hirlam hirlam 14160000 Nov 7 13:34 fc20101105_00+003
-rw-rw-r--. 1 hirlam hirlam 960000 Nov 7 13:34 fc20101105_00+003md
-rw-rw-r--. 1 hirlam hirlam 14160000 Nov 7 13:34 fc20101105_00+004
-rw-rw-r--. 1 hirlam hirlam 960000 Nov 7 13:34 fc20101105_00+004md
-rw-rw-r--. 1 hirlam hirlam 14184000 Nov 7 13:34 fc20101105_00+005
-rw-rw-r--. 1 hirlam hirlam 960000 Nov 7 13:34 fc20101105_00+005md
-rw-rw-r--. 1 hirlam hirlam 14184000 Nov 7 13:34 fc20101105_00+006
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+006md
-rw-rw-r--. 1 hirlam hirlam 14172000 Nov 7 13:34 fc20101105_00+007
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+007md
-rw-rw-r--. 1 hirlam hirlam 14136000 Nov 7 13:34 fc20101105_00+008
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+008md
-rw-rw-r--. 1 hirlam hirlam 14136000 Nov 7 13:34 fc20101105_00+009
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+009md
-rw-rw-r--. 1 hirlam hirlam 14172000 Nov 7 13:34 fc20101105_00+010
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+010md
-rw-rw-r--. 1 hirlam hirlam 14172000 Nov 7 13:34 fc20101105_00+011
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+011md
-rw-rw-r--. 1 hirlam hirlam 14184000 Nov 7 13:34 fc20101105_00+012
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+012md
-rw-rw-r--. 1 hirlam hirlam 14184000 Nov 7 13:34 fc20101105_00+013
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+013md
-rw-rw-r--. 1 hirlam hirlam 14148000 Nov 7 13:34 fc20101105_00+014
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+014md
-rw-rw-r--. 1 hirlam hirlam 14148000 Nov 7 13:34 fc20101105_00+015
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+015md
-rw-rw-r--. 1 hirlam hirlam 14148000 Nov 7 13:34 fc20101105_00+016
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+016md
-rw-rw-r--. 1 hirlam hirlam 14172000 Nov 7 13:34 fc20101105_00+017
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+017md
-rw-rw-r--. 1 hirlam hirlam 14172000 Nov 7 13:34 fc20101105_00+018
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+018md
-rw-rw-r--. 1 hirlam hirlam 14172000 Nov 7 13:34 fc20101105_00+019
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+019md
-rw-rw-r--. 1 hirlam hirlam 14172000 Nov 7 13:34 fc20101105_00+020
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+020md
-rw-rw-r--. 1 hirlam hirlam 14160000 Nov 7 13:34 fc20101105_00+021
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+021md
-rw-rw-r--. 1 hirlam hirlam 14148000 Nov 7 13:34 fc20101105_00+022
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+022md
-rw-rw-r--. 1 hirlam hirlam 14148000 Nov 7 13:34 fc20101105_00+023
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+023md
-rw-rw-r--. 1 hirlam hirlam 14184000 Nov 7 13:34 fc20101105_00+024
-rw-rw-r--. 1 hirlam hirlam 972000 Nov 7 13:34 fc20101105_00+024md
-rw-rw-r--. 1 hirlam hirlam 50 Nov 7 13:34 history
-rw-rw-r--. 1 hirlam hirlam 4534771 Nov 7 13:34 HL_Cycle_2010110500.html
-rw-rw-r--. 1 hirlam hirlam 149304 Nov 7 13:34 HL_MakeCycleInput_2010110500.html
-rw-rw-r--. 1 hirlam hirlam 3360000 Nov 7 13:34 sfcan
-rw-rw-r--. 1 hirlam hirlam 2 Nov 7 13:34 sfcan_done
-rw-rw-r--. 1 hirlam hirlam 704284 Nov 7 13:34 HL_PPCycle_2010110500.html
-

Re:Running a HIRLAM model with internal nesting 8 years 2 months ago #449

Also i checked each for kpds5=105:kpds6=105:kpds7=0:
this filed available in all fc*md and fc* files

Re:Running a HIRLAM model with internal nesting 8 years 2 months ago #450

  • Eoin Whelan
  • Eoin Whelan's Avatar
  • OFFLINE
  • Gold Boarder
  • Posts: 195
  • Thank you received: 34
The files matching the pattern fcYYYYMMDD_HH+0[0-9][0-9] are used, eg fc20101105_00+003. These are model level files and are used as boundaries.

The "md" files are model diagnostic files.

The "ve" files are used for verification.

Re:Running a HIRLAM model with internal nesting 8 years 2 months ago #451

may be i didn't correctly specify the path for boundry files in Env_input

my boundry files are located :
/home/hirlam/scratch/hl_arc/exp_rshu/2010/11/05/00

where exp_rshu the name of exp with main domain which was run for DTG=2010110500 LL=24

Re:Running a HIRLAM model with internal nesting 8 years 2 months ago #456

how can i set $INBDEXP at ECMWF for nesting if i'm using ec server ???

Re:Running a HIRLAM model with internal nesting 8 years 2 months ago #457

  • Eoin Whelan
  • Eoin Whelan's Avatar
  • OFFLINE
  • Gold Boarder
  • Posts: 195
  • Thank you received: 34
Hi Suleiman

I am running a test on ecgate at the moment using RCRa as the driving model. My experiment is still generating the climate file so I cannot say whether it will be successful or not.

I set
#
INBDDIR=ec:/hirlam/RCRa
#
in scripts/Env_input.

Someone else may have different ideas. I will be in touch when I have some results.

Eoin

Re:Running a HIRLAM model with internal nesting 8 years 2 months ago #458

  • Eoin Whelan
  • Eoin Whelan's Avatar
  • OFFLINE
  • Gold Boarder
  • Posts: 195
  • Thank you received: 34
This seems to have worked for me. Have a look at ecgate:/home/ms/ie/dui/hl_home/testNestHHH

Let me know if you have any issues accessing this directory.

Eoin

Re:Running a HIRLAM model with internal nesting 8 years 1 month ago #487

Eoin: I have sent you an email
unfortunately i couldn't access to you ecgate:/home/ms/ie/dui/hl_home/testNestHHH/scripts directory :(

Re:Running a HIRLAM model with internal nesting 8 years 1 month ago #488

  • Eoin Whelan
  • Eoin Whelan's Avatar
  • OFFLINE
  • Gold Boarder
  • Posts: 195
  • Thank you received: 34
Hi Suleiman,

I have made the scripts directory readable.

I will have a look at your experiment setups mentioned in your e-mail next week.

Eoin
  • Page:
  • 1
  • 2
Time to create page: 0.108 seconds