Model domain for SMOKE and WRF

SMOKE
1

Hi all,

I am attempting to run SMOKE 4.7 with the goal of generating inline plume rise input for CMAQ. I have configured the domain for a 12km WRF to 472 columns by 312 rows, which extends beyond the area set for SMOKE (459 by 299). What would happen if I run SMOKE with its predefined area for point sources and use my extended domain WRF output files for insertion into CMAQ?

2

Hello,

There should not be an issue using the inline point emissions files from a smaller domain as long as the projection is the same, including origins and standard parallels. Although this is not an issue from the emissions side, you may want to investigate if nesting of the WRF domain is appropriate for your CMAQ scenario.

1 Like
3

I would also note that it is customary practice to discard a ring of several WRF boundary cells when processing them (typically through MCIP) for running CMAQ. For example, the typical “BTRIM = 5” setting in MCIP would take a wrfout file with these dimensions

    west_east = 471 ;
    south_north = 311 ;
..
    west_east_stag = 472 ;
    south_north_stag = 312 ;

and process files for a CMAQ grid with 459 columns and 299 rows.

2 Likes
4

Thank you for your answers.

So, do I need to consider this ring of several boundary cells? I mean I also ran WRF for 2 nesting domain (301 by 301 and 77 by 77), for domain setting in SMOKE I need to define gridding that extend beyond these areas?

5

I am not an expert on this, but yes, in my understanding you generally do not want to use the maximum-possible extent of the WRF domain to drive CMAQ because meteorological fields in the outermost WRF grids may be compromised by “boundary effects” - others with a better understanding of WRF modeling and dynamics can hopefully provide a more detailed explanation.

In the MCIP run script, the following comment discusses this:

#-----------------------------------------------------------------------
# Set number of meteorology "boundary" points to remove on each of four
# horizontal sides of MCIP domain.  This affects the output MCIP domain
# dimensions by reducing meteorology domain by 2*BTRIM + 2*NTHIK + 1,
# where NTHIK is the lateral boundary thickness (in BDY files), and the
# extra point reflects conversion from grid points (dot points) to grid
# cells (cross points).  Setting BTRIM = 0 will use maximum of input
# meteorology.  To remove MM5 [WRF] lateral boundaries, set BTRIM = 5.

In the case of a wrfout file with a west_east dimension of 471 and a south_north dimension of 311, the maximum CMAQ domain that could be defined setting BTRIM = 0 in MCIP would be 469 columns and 309 rows, because a 1-cell perimeter is needed to generate the METBDY files with a standard NTHIK setting of 1.

In CMAQ, the modeling domain is defined using the GRIDDESC file generated by MCIP, which depends on your BTRIM settings. In general, it is easiest if all gridded input files (emissions, meteorology, etc.) provided to CMAQ are on that same grid. In no case can they be on a smaller grid than that defined by GRIDDESC. They can be windowed from a larger grid, but that requires care to make sure that all windowing conditions are satisfied.

As noted above by @james.beidler, inline point source emission files can be processed for a larger or smaller domain than that used in the CMAQ run, as long as the projection parameters used during point source file processing match those defined in the GRIDDESC file used by CMAQ.

1 Like
6

Thank you both @james.beidler & @hogrefe.christian