Creating a consistent CONUS grid

[dmv_fsf]

In my project I am trying to create a consistent concentration grid that covers CONUS. I will have emissions files for different fire events and want to be able to perform statistics on the results.

For example, I have one fire event happening at (lat lon) = (36.1419 -118.1856)
In the CONTROL, I set the center to be the approximate location of CONUS (center-lat center-lon) = (38.9 -98.5)
I also set the span to be wide enough for CONUS: (span-lat, span-lon) = (40, 60).

So from this setup, I expect HYSPLIT to create a concentration grid where the top-most latitude is 38.9 + 20 = 58.9 degs lat (and then similarly for the other edges). However, the resulting grid stays over the emission location, and does not cover all of CONUS (see image attached).

I have tried using ridiculous values for the span in the CONTROL file like (140, 160) and (1400, 1600) but the final grid remains the same, the HYSPLIT runtime stays the same, and so does the concentration output binary file size.

Is there some combination of emissions file and CONTROL file settings that I would have to do to create a consistent concentration grid over CONUS, but also use emissions files for fire events occurring anywhere and at anytime in CONUS?

PS: On my metdata
You will notice that I am using "NARR_large_file2" which is 10 years of NARR data which I have "cat" 'ed together to make the script easier to manage for my use case.

[alicec]

You are doing everything right.

The concentration output (cdump) file stores information about the grid in the header.
The information in the header that specifies the grid is
number of latitude points,
number of longitude points,
grid spacing (delta-lat, delta-lon)
grid lower left corner (lat, lon),
number of vertical levels,
height of each vertical level.

As long as that information is the same in each of your cdump files, then they can be combined.

Non-zero concentrations are not stored in the cdump files.
The concentrations are stored with their value and the indices of the grid square that the belong to.
The indices are converted to lat-lon points by the plotting (and other programs) using the grid specifications
found in the header.
The ramifications for this are
1. Cdump File size does not depend on how size of the concentration grid you defined. It only depends on the number of grid squares which have above zero concentrations.
2. The plotting programs will automatically 'crop' around the area of the plume.

Try creating another run with the same grid and a location way over on the east coast or something. Then try combining them.

More information on cdump file format.
https://www.ready.noaa.gov/hysplitusersguide/S363.htm

[dmv_fsf]

Thank you for the explanation. I went ahead and dug into the header of my cdump files and re-ran HYSPLIT changing only the span in each. After doing so I was able to see what you were talking about (in the red box in the image attached). I am now going to try combining an east coast emission with this one.

[dmv_fsf]

Doing the HYSPLIT run over a lat-lon in New York City resulted in the first image.
The second image was after using `conmerge` to combine the two files.

Thank you again for the clarification.