Some severe fires, such as the Peirces Creek Fire of April 1991, are driven by surface winds, which push the fire across the ground. In fact at that fire the tops of some pines were scorched and frozen in a horizontal position, like wind vanes. This indicates a very strong wind and a very brief peak heat pulse. Other severe fires are driven by vertical motion of air - i.e. the air's instability.
To quote from a research paper...(see Note 1)
Holding fuel type, topography, and all other meteorological factors constant, atmospheric instability promotes the spread and intensity of fires by increasing the following:
For the technically minded, the index reflects the temperature difference between heights were the air pressure is 850 and 700 hectopascals; and also the dew point depression at the 850 hectopascal height. Each of these are scored from 1 to 3, and the two are added together (thus the range of 2 to 6 for the Haines Index).
The finding in Tasmania suggest that using Haines as well as McArthur gives a better readiness indicator. In the ACT we have had no chances to test whether the same holds true - in the years that we have had Haines Indices forecast we have had no suitable fires.
This is good.
So whenever the Haines Index is a six, or maybe a five, Fire Controllers should be aware of the potential for the fire to grow well beyond what would be expected from the McArthur Indices. A large, vigorous convection column could be dragging down very dry air from aloft to replace the air moving upwards. If a large cumulus cloud forms in the convection column of a large fire, watch out, as evaporation of moisture from the cloud can generate a sudden downwash of air that can make a fire explode.
Note 1: "The Evaluation of Idaho Wildfire Growth using the Haines
Index" by Paul Werth & Richard Ochoa 1993.
From the 'Weather and Forcasting' Journal Vol. 8 pages 233-234.