Bug in applying adjustment factors by representative county in Movesmrg.

[hnqtran]

This bug is part of the issues described in the ticket #104.

When applying adjustment factors by representative county (USE_REF_CONTROL_FAC_YN = Y) in Movesmrg: if there is a group in which a rep county is mapped to several counties, where the rep county itself has no activity (which can happen with hoteling / RPH in particular) but other counties in the rep county group do have activity, then the adjustments are not applied to the other counties in the rep county group.

[callen52]

Here are some example SMOKE-MOVES inputs to test with. I recommend testing with RPH. You can take most other onroad inputs that you would need for testing from 2022 platform (https://gaftp.epa.gov/air/emismod/2022/v1/)

HOTELING_2022v1_2026hc_MOVES4_25jul2024_nf_v1.csv - This is hoteling activity for 2026 (but can be used with the posted inputs from 2022 for testing purposes). Note that county 55075 has no activity in this dataset, but other counties in the 55075 reference county group do have activity, e.g. 55005.

cfpro_2026hc_MOVES4_OTAQrules_annual_RPH_26jul2024_v0.csv - Here is a CFPRO file we have been using for 2026, which applies factors by reference county (USE_REF_CONTROL_FAC_YN = Y). County 55075 is a reference county and is listed here. But what I am finding it that since 55075 itself has no hoteling activity, these factors are not applied to any of the counties in the ref county group which do have hoteling activity (e.g. in county 55005).

HOTELING_2022v1_dummy_bug_workaround_30jul2024_v0.csv - This is an additional hoteling activity dataset which I included in my SMOKE-MOVES run as a workaround, and which includes a very tiny amount of activity in county 55075, among other ref counties with this issue. When I include this in the SMOKE-MOVES run, the CFPRO factors for 55075 are properly applied to the entire ref county group.

Monthly_onroad_RPH_12US1_2026hc_cb6_22m_20240730111852.csh.txt - Our run script, in case there are any questions about parameter settings.

[bokhaeng]

@callen52 I have reviewed your sample files and identified the cause of this issue. In reality, there is no bug in SMOKE for this issue since SMOKE is designed to drive any emission processing based on the list of inventory sources sorted by Smkinven. If there is no inventory sources from this reference county (e.g., 55075) listed in the inventory file, then SMOKE is intentionally designed to ignore any following processing like grdmat, spcmat, temporal, and Movesmrg.

To artifically add a new inventory source without any activity (e.g., VMT, POP, HOTELLING,,), we need to add an option to read the MCXREF (reference county x-reference) input file and this mssing reference county inv sources if they are not existed in the original inventory file. It can not be a default option but limited to mobile sources only.

My question is whether this option is needed or not? Better to add these dummy zero-valued inventory sources or create a new option to Smkinven? Without this option, the file size will be slightly increased by no of lines for those missing inventory sources.