The fate of smoke billowing from industrial stacks, like those shown here, is a common concern — not least to those who have just hung out their washing. Writing in Atmospheric Environment (36, 4603–4615; 2002), R. W. Macdonald and colleagues describe modelling investigations of how smokestack configuration with respect to the prevailing wind affects the behaviour of smoke plumes.

As a single hot plume rises, vortices form within it and cool air is drawn in, so reducing plume buoyancy until it rises no further. But when two plumes merge, less cool air is entrained — because of the lower surface-area-to-volume ratio — and the plume is buoyant for longer, rising higher. The distance between smokestacks clearly influences plume merging. But what about the arrangement of the stacks?

Macdonald et al. used a water flume, studying hot-water plumes rising from two 10-cm 'stacks' by repeatedly dragging a grid of temperature probes through the plumes. Plumes in line with the overall flow in the flume quickly mixed and rose higher, as expected: the first plume shields the second from the flow, so that the latter bends less and rises into the first in a way that allows the internal vortices to mingle without destroying overall plume integrity.

Plumes from stacks aligned across the flow did not mix, and rose more slowly — mixing being hindered because the vortices at the edges of each plume tend to oppose each other. Interaction of vortices on the plume edges can also create a downwash effect, perhaps explaining the slowed plume rise.

Thus, smoke from chimneys set in line with wind direction, rather than across it, is likely to rise higher, travel further, and presumably become more dispersed and diluted before reaching the ground.