The SMOKE ORL formats use code numbers (usually Chemical Abstracts Service [CAS] numbers) to distinguish which chemical compound or inventory pollutant is provided on each line of the file. Smkinven matches these numbers with the CAS numbers from an inventory table (INVTABLE) file, described in Section 8.10.3, “INVTABLE: Inventory table”. The CAS number does not necessarily have to be a valid CAS number; it can be any number as long as there is a match between the numbers in the SMOKE ORL file and the inventory table. The inventory table provides the inventory data names, such as the pollutant names that SMOKE uses in the remaining processing steps. Note that the SMOKE inventory pollutants may not be identical to the pollutants in the inventory fed to SMOKE because of the aggregation/disaggregation that is performed by Smkinven (see Section 2.9.7, “Aggregate or disaggregate toxics emissions” for more information).

For toxics processing, if multiple inventory data names apply for the same CAS number, the Factor column of INVTABLE will contain the split factor used by Smkinven to disaggregate the emissions from that CAS number to multiple inventory data values. If multiple CAS numbers apply for the same pollutant name, then Smkinven will sum these emissions, but will not report duplicate records unless there are indeed duplicates in the inventory file. This is described more fully in Section 2.9.7, “Aggregate or disaggregate toxics emissions”. Duplicate reporting is described more in Section 2.9.4, “Check for duplicate records”.

The use of the inventory data name as the unique pollutant identifier in SMOKE differs from the approach of EMS-HAP, in which the SAROAD code is the unique identifier for the pollutants to be modeled. Because we anticipated that some toxics pollutants that do not have unique SAROAD codes (e.g., divalent particulate mercury) would need to be modeled explicitly, we did not want to take this approach. If we had, the user would have been required to create fake and unique SAROAD codes to be able to model these emissions explicitly.