[WRF-Chem 4.x] Namelist parsing failure in &domains section (time_step issue) leading to MPI abort in LCZ simulation

[bhhhh]

I’m encountering a fatal namelist parsing error when running WRF-Chem (configured with the LCZ urban canopy scheme) for my undergraduate thesis on urban heat island and ozone effects. The model crashes immediately at startup, with the following error in rsl.error.0000:
Maybe here?: sf_urban_physics = 1, 1
Maybe here?: use_wudapt_lcz = 1, 1
------ ERROR while reading namelist domains ------
Maybe here?: time_step = 90, 30
Maybe here?: time_step_fract_num = 0, 0
-------------- FATAL CALLED ---------------
FATAL CALLED FROM FILE: <stdin> LINE: 12500
ERRORS while reading one or more namelists from namelist.input.
-------------------------------------------
application called MPI_Abort(MPI_COMM_WORLD, 1) - process 0

Questions:​

1. Could the GCC 4.x version be incompatible with WRF-Chem 4.x, leading to this namelist parsing failure?
2. Are there any specific namelist syntax rules or compatibility considerations I’ve missed when using the LCZ scheme with double nesting?

Environment details:​

• WRF-Chem version: 4.x (configured with LCZ scheme)
  Compiler: GCC 4.x (I’m wondering if this could be a factor)

Thank you so much for your help and suggestions in advance！

 

[kwerner]

Hi, As your error states, the issue is with some of your namelist settings. It looks like you've added column 2 settings for many namelist parameters that only allow a single setting. For e.g., use_wudapt_lcz only allows a single entry. So instead of "use_wudapt_lcz = 1, 1" you should set "use_wudapt_lcz = 1"

There are many others in your namelist that have this problem. The best advice is to use a default namelist.input file, and if the variable only has a single setting in the default namelist, they you shouldn't add additional columns. Because many of your settings aren't in the default namelist, you'll need to go to the WRF/Registry directory and then search for each variable in the registry files to determine if it requires multiple settings (you'll see "max_domains" listed in the variable line), or just a single setting.

For e.g.,
"max_domains" in this variable entry tells you that you need one setting per domain.

rconfig   integer   e_we    namelist,domains   max_domains  32      irh    "e_we"          ""      ""

But the time_step variable only allows a single entry - not an entry for each domain. Instead of max_domains, it simply has a "1" in that column.

rconfig   integer   time_step      namelist,domains      1          -1       ih   "time_step"