Good question: this is the first I've seen of ARL running particle cross section analysis right at the start of the controlled fire before the mean wind shifted. So at the initial burst, they're modeling (using I suspect a different model than the one I used to calculate the forward trajectory) the most likely dispersion integrating estimated particle release per hour. That's also starting a full 12 hours earlier than mine, right as the trough is moving through, more on that below.

What this shows is particles initially contained in the lowest 3,000 ft of the atmosphere that advect NE, then N as they're mixed higher into the atmosphere (shades of blue). This likely is along the first upper-level trough to pivot through here last week, though these particles being an issue here seems almost non-existent given what we know about the lifespan of vinyl chloride gas & how high up these particles were mixed. Then as the winds became more predominately WNW (second image), the plume orients ESE, with concentrations contained in a smaller area than I think some feared would be otherwise.