Intervals between prescribed fires in Australia: what intrinsic variation should apply?

Because of increasing concern over the constancy of intervals between prescribed fires within a vegetation type, we examine various sources of evidence that can be used to determine variation appropriate to the conservation of biodiversity while minimizing the chances of economically destructive fires. Primary juvenile periods of plants (especially of 'serotinous seeders') and non-breeding periods of birds (especially poorly dispersed species) suggest extreme lower limits for fire intervals whereas longevity of plant species which usually only reproduce after fire, set the extreme upper li:mits. Modelling of the behaviour of selected plant and animal species may be used to set 'optimal' mean intervals. Historical fire-interval data might seem a useful way to determine the variation about the mean fire-interval but data are scarce and interpretations are controversial. The Weibull distribution and its special case, the negative exponential distribution, have been the most supported in North American studies of unplanned fires. It has been argued that fire-interval distributions, before European settlement at least, were largely the result of large fires during, or following, extreme weather events (dry in forests, wet in the arid zone). Long weather records are most beneficial when they can be related to the areas burned each year. Practical solutions to the question 'what range of fire intervals should be used at any one site' may be achieved using highly simplified skewed distributions, constructed on the basis of land-management objectives. (C) 1998 Elsevier Science Ltd. All rights reserved.