The behaviour of buoyant plumes in neutral and stable conditions in canada

Abstract

Plume dispersion parameters obtained during six years of research by the Canadian Combustion Research Laboratory (CCRL) have been evaluated and compared to the standard predictive relationships established by Briggs and Pasquill-Gifford. The data correspond to neutral and stable conditions, stability class C to F but excludes any limited-mixing or layered atmosphere studies or situations where topography influences plume spread. In all, ten sources in five geographic regions were studied during Spring, Fall and Winter. It has been shown that the plume rise data can be represented by the Briggs and Moore form of relationship for neutral conditions. The data suggest a proportionality constant of 0.87 and a levelling-off at 15 stack heights when constrained to the 2 3 power law required by the Briggs continuous model. The unconstrained regression yielded a constant of 0.5 and 0.71 power law relationship with downwind distance, between the Briggs and Moore relationships. However, both of the latter significantly overestimate the measured data. In stable conditions the Briggs levelling-off value of 2.9 for the dimensionless plume rise is in good agreement with the findings in this paper, but this occurs at a dimensionless downwind distance of 18.4 rather than 2.4. A slight variation in the maximum plume rise with the bulk Richardson number over the dispersion zone has been noted for stable conditions. The Pasquill-Gifford plume spread curves as modified by Bowne for rural conditions represent a convenient means of comparing the data obtained under corresponding stability and topographic conditions in Canada. The variation of measured horizontal spread parameters, σ y with downwind distance differed significantly from the P/G curves, being wider by at least two stability classes for unstable/neutral, i.e. A/B rather than C/D, and for stable, C rather than E. In the case of very stable F class, the measured σ y values corresponded to A class, six classes wider than expected. The vertical spread parameters, σ z were usually in agreement with P/G curves from 4 to 10 km from the source, closer to the source σ z was greater and farther from the source σ z was less than predicted. It was also found that the bulk Richardson number could be used to classify the plume spread parameters in a similar manner to the P/G stability classes.