WRF chem simulations restart issues

[Leo Matak]

Dear all,

I am trying to run longer periods of simulations (5 days), and it is very difficult to do it all in one go. So what I would like to do is to split the simulation using the restart files. And the restart works if the restart timers are set at each 12h. However, when running on Derecho HPC the wall clock time limit is also 12h which means I might not be able to do my 12h of simulation in 12 h of wall clock time. So instead I change the restart timer to each 10h. However, I cannot make it work. I tried many things, and analyzed the error:

!this is the only one that breaks! 

 io_form_auxinput5                    =2,

 auxinput5_inname                    = 'wrfchemi_<00>z_d<domain>',

 auxinput5_interval_m                = 60,

 frames_per_auxinput5                = 12,

 auxinput6_inname                    = 'wrfbiochemi_d<domain>',

 io_form_auxinput6                   = 2,

the io_form_auxinput5 is the issue. If I comment that one out, it runs. 

If I don't comment it out the error I get is:

mediation_integrate: med_read_wrf_chem_emissions: Read emissions for time 2020-02-05_10:00:00

d01 2020-02-05_10:00:00  Error trying to read metadata

d01 2020-02-05_10:00:00  Input data is acceptable to use:

           0  input_wrf: wrf_get_next_time current_date: 2020-02-05_10:00:00 Status =         -102

d01 2020-02-05_10:00:00  ---- ERROR: Could not find matching time in input file

NOTE:       1 namelist vs input data inconsistencies found.

-------------- FATAL CALLED ---------------

FATAL CALLED FROM FILE:  <stdin>  LINE:    1303

NOTE:  Please check and reset these options

Finally, I would also like to point out that I've changed the Times variable in all the wrfchemi* files to: 

 Times =

  "2020-02-05_00:00:00",

  "2020-02-05_01:00:00",

  "2020-02-05_02:00:00",

  "2020-02-05_03:00:00",

  "2020-02-05_04:00:00",

  "2020-02-05_05:00:00",

  "2020-02-05_06:00:00",

  "2020-02-05_07:00:00",

  "2020-02-05_08:00:00",

  "2020-02-05_09:00:00",

  "2020-02-05_10:00:00",

  "2020-02-05_11:00:00" ;

Please let me know if there is any workaround this. Also it might be worth pointing that untill the fifth day the restart timers were at 12h, but then after fifth day I change them to 10h. 

Thank you,

Leo

[Gabriele Pfister]

Hi Leo

I have not run into this issue but my guess is that any run has to start at one of the initial times in the wrfchemi files. In the case of using io_style_emissions=1, i.e. using two 12-hour emission files the run might have to start at either one of the initial times in the emissions files. This would be 00 UTC or 12 UTC. One way around this would be to create hourly emission files (io_style_emissions=2) from the two 12-hour files. 

Hope this helps

Gabi

--

================================
Gabriele Pfister

Atmospheric Chemistry Observations & Modeling (ACOM) Lab

NSF National Center for Atmospheric Research (NCAR)

Email: pfi...@ucar.edu

Temporary Home Office Phone: +1 303 241 1329

Work Phone: +1 303 497 2915

Web: https://staff.ucar.edu/users/pfister