This README contains brief instructions on how to run SMOKE for source apportionment applications. EPA performed a source apportionment run based on the 2017eh_cb6v2_v6_11g case, called 2017eh_ussa_cb6v2_v6_11g ("ussa" = US source apportionment), for air quality modeling with the CAMx model. The general procedure for creating model-ready emissions files for source apportionment is as follows, with further detail below: - Run SMOKE for each sector in source apportionment mode, creating model-ready emissions in CMAQ inline format - Convert the CMAQ-ready emissions files to CAMx format Details on running SMOKE in source apportionment mode: 1) Create one or more "source groups" files, which lists the "tags" assigned ] to each source. The format of the source groups file is outlined in the SMOKE documentation: https://www.cmascenter.org/smoke/documentation/3.6.5/html/ch08s10s12.html For 2017eh_ussa_cb6v2_v6_11g, source apportionment tags were assigned by state. Each US state received its own tag for all anthropogenic emissions (excluding wild and prescribed fires, biogenics, and particulate dust). Biogenic and particulate dust (afdust) emissions received a separate tag; all Canadian and Mexican emissions received another separate tag; offshore marine emissions received their own tag; and wild and prescribed fires received their own tag as well. The 2017eh_ussa case used two SOURCE_GROUPS files: one for all US sectors including the state-specific tags, and one for Canada and Mexico including the Canada+Mexico tag. Tags for afdust, ptfire, and beis were applied outside of SMOKE (see item #3). These SOURCE_GROUPS files are included in smoke_scripts_utils/2017eh_ussa_source_apportionment_scripts_inputs.zip: smkmerge_groups_2011ed_ussa_US_02apr2015_v1 smkmerge_groups_2017eh_ussa_nonUS_02apr2015_v0 2) Start with the "base case" (e.g. 2017eh) SMOKE run scripts. Two additional variables/inputs must be defined in order to run SMOKE in source apportionment mode: - setenv SMK_SRCGROUP_OUTPUT_YN "Y" # triggers source apportionment mode - setenv SOURCE_GROUPS [name of source groups file] 3) Run each sector through SMOKE. For source apportionment, SMOKE creates "inline" format emissions files for all sectors, even area and mobile source sectors. These inline emissions files, typically named sginln_mole_*, include all emissions in the sector, independent of whether the emissions are considered "low-level" or "elevated". In other words, for point sectors that normally have emissions split into a low-level file and an elevated file, the source apportionment "sginln" file includes both low-level and elevated emissions. These sginln files, along with the associated stack_groups_out files, can be used as inputs to CMAQ, but we convert them to CAMx format later on for our modeling. When running in source apportionment mode, SMOKE still creates the usual emis_mole_* and/or inln_mole_* emissions files, but those files are not needed for air quality modeling if running source apportionment. When running SMOKE, special considerations must be taken for the following sectors: AFDUST: The scripts which apply transportable fraction and meteorological adjustments to post-SMOKE afdust emissions files do not work with the "sginln" files that SMOKE creates for source apportionment. This sector includes PM emissions only, and the focus of the 2017eh_ussa model run was ozone production, for which PM emissions are not vital. Therefore, it was decided to not apply state-specific tags to afdust, and instead apply a sector-wide tag during the CAMx conversion step of the process. So instead of running afdust through SMOKE for source apportionment, we run SMOKE in "normal" 2-D mode, and then apply a sector-wide tag during the CAMx conversion step. OTHAFDUST: Same as afdust. PTFIRE: If we were modeling with CMAQ, we could run ptfire through SMOKE with source apportionment just like the other sectors. However, because we are modeling with CAMx, we must have SMOKE create 3-D layered ptfire emissions. This is a problem because SMOKE only supports the creation of inline format emissions for source apportionment, not 3-D layered emissions for source apportionment. So, just like with afdust, we do not run ptfire through SMOKE with source apportionment. We run it in normal 3-D mode, and then apply a sector-wide tag during the CAMx conversion step. BEIS: Similar to afdust and ptfire, since this sector is only receiving a sector-wide tag, this sector was run normally through SMOKE, with a sector-wide tag applied during the conversion to CAMx format. ONROAD: Two things to consider with source apportionment for onroad: - Running SMOKE-MOVES with source apportionment requires a very large amount of memory, particularly for RPD processing. It may not be possible to run SMOKE-MOVES with source apportionment for the entire US on all systems. - The onroad "merge" step, which combines emissions from RPD, RPP, RPV, and RPH, is omitted when running onroad with source apportionment. The individual files from RPD, RPP, etc are input directly to CMAQ (if running CMAQ) or to the CAMx converter (if running CAMx). 4) For source apportionment cases, the sector merge (Mrggrid) should only merge sectors that are NOT getting source apportionment tags applied. If all sectors are getting tags - including sectors like afdust that are getting tagged outside of SMOKE - then Mrggrid does not need to be run at all. This was true in the 2017eh_ussa case. The source appportionment inline-format emissions (sginln_mole_*) themselves are not merged. The sector emissions are all input directly to CMAQ or the CAMx converter. Details on converting SMOKE outputs (CMAQ-ready inline files) to CAMx format for source apportionment: The source apportionment CAMx conversion requires the installation of the "base case" CAMx conversion executables and all ancillary files from 2011v2_all_speciations_camx_conversion.zip. The 3D wildfires and 3D prescribed fires must also be converted to a CAMx ready ptsr file prior to running the source apportionment conversion. Additional scripts for converting source apportionment emissions to CAMx format are in: smoke_scripts_utils/2017eh_ussa_source_apportionment_scripts_inputs.zip The scripts in this zip are *sample scripts only* and are not configured the same way as the scripts in other zips, in terms of directory names and paths. For a source apportionment CAMx run, all sectors are tagged and written as point or "ptsr" type emissions files. This is a multi-step process involving the conversion and tagging of certain low-level sources (beis, othafdust, and afdust) along with the conversion of tagged inline files from the remaining sectors. All low-level sources should be converted prior to running the "camx_convert_2017eh_ussa_12US2.csh" script, which merges the sector-specific ptsr files into a single CAMx ready emissions file. The seasalt emissions generation script should also be run following the completion of the merged CAMx ready emissions files.