I am trying to run an annual pt script using the pre-compiled version of SMOKE (v5.1) and I am getting an error message that there is a problem in RDSREF in that there is both “COMBO” and fractional profile entries. I can run the script with no VOC emissions but it fails when I include the VOC emissions. I have tried numerous ways to try and find out exactly what the problem is but have not had any success.
I have included the emission file that is successful (*ADMF_short.txt), the emission files that fails (*ADMF.txt), the speciation log file for the failed case (*log.txt) as well as the common ancillary files used for both cases (gspro, gspro_combo, gsref and invtable).
Any suggestions on how to track down and solve the issue would be greatly appreciated. Thanks.
The quick fix is to edit the GSREF to remove all profile assignments for SCCs 2310011506, 2310021010, and 2310021506 that specify a US FIPS. This will leave just the overall defaults for those SCCs.
The issue likely occurs as a SMOKE limitation which prevents the use of the separate GSPRO_COMBO file (in which the GSREF says “COMBO” and the same set of fractions is applied to all COMBO SCCs) and fractional GSREF profiles (in which the same SCC or FIPS/SCC is listed multiple times in the GSREF with separate profiles, with fractional splits in GSREF Column M). Spcmat is coded to only be able to do one “combo” method or the other in a given sector, not both.
I’m not sure i understand exactly what is going on here so please bear with me.
When i remove the three sources from the emissions with the SCC codes you suggested were a problem the script runs to completion. However, when I look in the gsref file none of these sources have a “COMBO” entry to cause a problem so:
How did you select these sources and know they were the problem?
If i remove all references to the problem SSC codes in the gsref file and replace them with:
“2310011506”,“0000”,“VOC”,
“2310021010”,“0000”,“SO2”,
“2310021506”,“0000”,“VOC”…
i get the same error message. The gspro and gscn files have default “0000” entries for these species so I am expecting this to be the correct fix.
Any additional guidance you can provide on this matter would be greatly appreciated.
Sorry, just to be clear. The second method where I add the modified SCC codes to the gsref, with the original SCC codes removed, I get the same error message when I try processing the original emissions data.
Did the profile have an entry in the 13th column/column M? If so then it is automatically classified as a combo profile and cannot be used concurrently with those profiles classified as “COMBO” in the GSREF. This is true if any SCC has column M populated.
Many, many, many thanks for your advice. When I remove all cross-reference entries in the gsref file that have a non-blank entry in column M I am able to SMOKE the data. whoohoo!
A couple of follow up comments:
Error message thrown by RDSREF is really cryptic. In the next version of SMOKE I am not sure if there is any value in listing the offending SCC codes or some other information that would help to point to the issue. None of the offending cross-reference entries with non-blank column M mention “COMBO” so this really threw me off. In addition, none of the offending entries were tied to any counties that I was modelling so even if there was an offending entry in gsref I am not sure why SMOKE was trying to speciate emissions from this source?
I am not sure if there is any value to add to some more text to the SMOKE user manual regarding gsref. Column M is listed as optional but really could use a blurb to mention how/when to use it correctly.
Again, many, many thanks for your help getting me back on track.