The temporal variation of the days in a week is the same from week to week within a month because SMOKE uses weekly profiles. Also, it is common that the weekly profiles do not vary on Monday through Friday. Consequently, it is often desirable for long (e.g., annual) simulations to use a Monday-weekday-Saturday-Sunday (MWSS) approach. With this approach, SMOKE computes emissions for a representative Monday, weekday, Saturday, and Sunday within each month. The representative days cannot be the first day of the month (to prevent effects from the previous month from being included in the emissions data), a holiday (as set in the HOLIDAYS file), or the day after the holiday. Monday is distinguished from other weekdays because in multi-time zone cases in the Western Hemisphere with OUTZONE set to 0 (i.e. GMT), a few late-night Sunday hours are included in the hours at the start of the Monday file. In addition, one must specifically process the holidays and the day after holidays as separate runs for all holidays set by the HOLIDAYS file. During merging of emissions, the Monday, weekday, Saturday, and Sunday files are reused to create model-ready emissions for every day and hour needed. The holiday and day-after files are used for the holiday and day-after dates only.

This approach relies heavily on scripting to select dates for each month’s representative days and to ensure that the days that are run are consistent with the dates in the HOLIDAYS file. Scripting is also responsible for ensuring that the correct files are merged together by Smkmerge to create the model-ready files for all days of the episode.

There are some situations in which you cannot use the MWSS approach. In particular, on-road processing with MOVES must be run for all days of the episode, although some steps can be sped up by using the meteorology averaging approach. Additionally, biogenic emissions processing obtains all of its temporal variation from the meteorology data, and therefore must be run for all days. Fortunately, biogenic emissions processing is very fast compared to processing for other source categories.