[Wrf-users] WRF does not save output for more than 1 hour
Wang, Yaoping
wang.3866 at buckeyemail.osu.edu
Mon Nov 21 13:35:25 MST 2016
Hi All,
I ran into a weird problem running WRF. As I set it up, the history interval is every 60 minutes, and 24 time slices should be in each output file. The model behaved as expected until about 8 years into a long integration. Then the model does not save more than 1 time step. That is, when it restarted from day 2068-07-07 00:00:00, only 1 time steps exist in the file:
Times =
"2068-07-07_01:00:00",
whereas an ordinary history file should look like
Times =
"2068-07-06_01:00:00",
"2068-07-06_02:00:00",
"2068-07-06_03:00:00",
"2068-07-06_04:00:00",
"2068-07-06_05:00:00",
"2068-07-06_06:00:00",
"2068-07-06_07:00:00",
"2068-07-06_08:00:00",
"2068-07-06_09:00:00",
"2068-07-06_10:00:00",
"2068-07-06_11:00:00",
"2068-07-06_12:00:00",
"2068-07-06_13:00:00",
"2068-07-06_14:00:00",
"2068-07-06_15:00:00",
"2068-07-06_16:00:00",
"2068-07-06_17:00:00",
"2068-07-06_18:00:00",
"2068-07-06_19:00:00",
"2068-07-06_20:00:00",
"2068-07-06_21:00:00",
"2068-07-06_22:00:00",
"2068-07-06_23:00:00",
"2068-07-07_00:00:00" ;
I kind of got through day 2068-07-07 by setting history_interval = 120 in the namelist.input file. The model ran and gave the output times like :
Times =
"2068-07-07_01:01:08",
"2068-07-07_02:01:12",
"2068-07-07_03:01:12",
"2068-07-07_04:01:12",
"2068-07-07_05:01:12",
"2068-07-07_06:01:12",
"2068-07-07_07:01:12",
"2068-07-07_08:01:12",
"2068-07-07_09:01:12",
"2068-07-07_10:01:12",
"2068-07-07_11:01:12",
"2068-07-07_12:01:12",
"2068-07-07_13:01:12",
"2068-07-07_14:01:12",
"2068-07-07_15:01:12",
"2068-07-07_16:01:12",
"2068-07-07_17:01:12",
"2068-07-07_18:01:12",
"2068-07-07_19:01:12",
"2068-07-07_20:01:12",
"2068-07-07_21:01:12",
"2068-07-07_22:01:12",
"2068-07-07_23:01:12" ;
which makes no sense, because the history interval was 60 minutes-ish instead of the set 120 minutes. But at least it gave me a restart file on 2068-07-08 00:00:00. So I set the history interval back to 60 minutes and ran 2068-07-08, and again the model stopped after 1 hour:
Times =
"2068-07-08_00:00:00",
"2068-07-08_01:00:00" ;
A `grep csf rsl*` does not reveal any cfl error. The last lines of rsl.error.0000 look like:
Timing for main (dt= 72.00): time 2068-07-08_00:53:56 on domain 1: 0.50543 elapsed seconds
Timing for Writing wrfout_d01_2068-07-08_00:00:00 for domain 1: 0.41741 elapsed seconds
Timing for main (dt= 72.00): time 2068-07-08_00:55:08 on domain 1: 0.51300 elapsed seconds
Timing for Writing wrfout_d01_2068-07-08_00:00:00 for domain 1: 0.45703 elapsed seconds
Timing for main (dt= 72.00): time 2068-07-08_00:56:20 on domain 1: 0.66976 elapsed seconds
Timing for Writing wrfout_d01_2068-07-08_00:00:00 for domain 1: 0.40401 elapsed seconds
Timing for main (dt= 72.00): time 2068-07-08_00:57:32 on domain 1: 0.50574 elapsed seconds
mediation_integrate.b 1724 DATASET=HISTORY
mediation_integrate.b 1725 grid%id 1 grid%oid 2
Timing for Writing wrfout_d01_2068-07-08_00:00:00 for domain 1: 0.40513 elapsed seconds
Timing for main (dt= 72.00): time 2068-07-08_00:58:44 on domain 1: 0.50572 elapsed seconds
Timing for Writing wrfout_d01_2068-07-08_00:00:00 for domain 1: 0.95947 elapsed seconds
Timing for main (dt= 40.00): time 2068-07-08_00:59:24 on domain 1: 1.04635 elapsed seconds
Timing for Writing wrfout_d01_2068-07-08_00:00:00 for domain 1: 0.47012 elapsed seconds
Timing for main (dt= 35.88): time 2068-07-08_01:00:00 on domain 1: 0.55670 elapsed seconds
Timing for Writing wrfout_d01_2068-07-08_01:00:00 for domain 1: 0.41406 elapsed seconds
The namelist.input file is attached.
Does anyone have suggestions?
Thank you very much,
Yaoping Wang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/wrf-users/attachments/20161121/e5bcaed6/attachment.html
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: namelist.input.txt
Url: http://mailman.ucar.edu/pipermail/wrf-users/attachments/20161121/e5bcaed6/attachment.txt
More information about the Wrf-users
mailing list