Example: Smoke plumes from an industrial fire

Using radars to look at smoke plumes

On May 23, 1996, a large fire broke out in a paint factory on Laval Island northwest of downtown Montreal. The wind was blowing from the northwest and so the thick dark smoke plume that developed was blown over central Montreal. Unexpectedly, the plume was detected and tracked by the McGill's weather radar, wind profiler radar, and X-band Vertically Pointing Radar (VPR).

S-band radar image of plume This image shows echoes seen by the weather radar at 1:30 pm EST, about an hour and a half after the fire started. The yellow to red colour contours indicate the strength and location of the echoes seen by the radar at an altitude of 1.5 km. The echoes originated near the location of the fire and spread towards the southeast. The radar clearly is detecting something in the smoke plume from the fire, which is unexpected because smoke particles are too small to be seen by this type of radar. On this very old animation of the radar images, the radar, profiler, and fire positions are indicated by R,P,F; the clock indicates EST.


The echoes seen by the X-band vertically pointing radar are shown in this image. This radar is situated at the downtown campus of McGill next to the profiler on the roof of Burnside Hall and looks straight up into the atmosphere. The horizontal axis on the plot is time and the vertical axis is altitude. The contours show the strength of the echoes above the

VPR image of plume

radar as the plume drifted overhead. The green echoes at low levels are mainly due to insects. The brighter echoes are due to the smoke plume. Notice that the echoes have a spotty nature which indicates that we are seeing a small number of largish particles (probably ash), rather than a large number of tiny particles.

If we look at the dBZ scale we can see that the echoes are much weaker on this radar than they are on the weather radar. They are also much weaker than the echoes on the profiler (not shown). The wavelength of the weather radar and profiler are 10 cm and 33 cm respectively, but the wavelength of the X-band VPR is just 3 cm. By examining the strength of the echoes at the different wavelengths of the various radars, we were able to guess that the particles the radars are seeing are around 1 cm across.

Profiler image of plume

The profiler radar can measure the Doppler velocity of the echoes looking vertically, and this is plotted here. The blue and green contours indicate downward motion and the yellow to red colours indicate upward motion. Much of the time the contours are blue & green, which indicates that particles are falling at about 1 - 3 metres per second. There are also some orange contours above about 2 km; this probably indicates that particles at these levels are trapped by updrafts and convection at the top of the boundary layer as they are being blown along by the wind.

This study showed that radars can be used for studying smoke plumes. Weather radars can be used to track the direction of the plumes, and vertically pointing radars can be used to study the fallout of ash from the plumes. A fuller report on this study entitled "Radar Observations of a Major Industrial Fire" appears as the cover story in the May 1997 issue of the Bulletin of the American Meteorological Society.

Home      Facilities      Imagery      Who we are      Science      Data Archive