Table 3.

Plume rise model differences

Approach	Briggs	Sofiev	PBL500
Type	Empirical, analytical	Empirical, analytical	Statistical
Scheme	Plume height is a function of downwind distance, modified for wildfire by Pouliot et al. (2005); plume top is calculated, and the plume bottom is set to be 2/3 of plume top value	Energy-balance-based parameterisation (similar to convective cloud formulations) accounting for planetary boundary layer (PBL) height, power law dependence of fire intensity, and stability above the PBL, with four fitted tuneable parameters to match observed plume heights by MISR	Observation-based, consistent statistical approximation
Input parameters	Area burned, fuel loading, wind speed, duration of fire, heat content, fire location coordinates	Fire Radiative Power (FRP), potential temperature as a function of geometric height to derive Brunt–Väisälä frequency, PBL height, fire location coordinates	PBL height, fire location coordinates
Output parameters	Top and bottom of smoke plume	Top of smoke plume	Top of smoke plume
Previous comparisons with satellite observations	Tendency to underpredict (Raffuse et al. 2012); plume tops are generally higher than Sofiev owing to inherent incompatibility and formulation (designed for stack heights), with fires (Sofiev et al. 2012)	Poor to moderate; comparable with or better than a one-dimensional plume rise model (Sofiev et al. 2012; Paugam et al. 2016)	Not available
Ease of implementation in CMAQ	Currently used	Low to moderate	Very low
Existing implementation	SMOKE, CMAQ, HYSPLIT	CMAQ (by Baldassarre et al. 2015)	Not currently used