We have a new paper out in this week's Proceedings of the National Academy of Sciences Early Edition.  The paper suggests that subsistence related burning increases local landscape heterogeneity and may promote biodiversity in Australia's Western Desert.  What's really interesting about this is that promoting biodiversity is not the goal of individual hunters – they are really just trying to maximize their foraging returns.  My Stanford Anthropology colleagues Rebecca Bliege Bird and Doug Bird have been working with Martu foragers in Western Australia since 2000.  They gather amazingly detailed quantitative ethnographic and ecological data, focusing on such classic problems as understanding subsistence economy in a foraging population and the ecological factors favoring a sexual division of labor.

Martu foragers burn the climax spinifex grass in search of goanna lizards and in so-doing, promote the growth of plant species that cannot compete with adult spinifex, which is a nasty grass with high-silica content:

Spinifex is extremely flammable and once a fire is started in a patch, it will race through until it is pretty much all burned.  The vegetation that characterizes other successional stages is far less flammable.  Previous work by Doug and Rebecca has shown that burning increases hunters' encounter rates and probability of capturing goanna lizards and other prey species. 

Using their detailed information on hunting and burning, we correlated foraging activity with measures of habitat heterogeneity derived from satellite imagery.  We compared local habitat heterogeneity in areas where Martu were actively foraging (and burning) to areas where there was no foraging and the fire regime was instead driven by lightning ignitions.  What we found was that the heterogeneity in the foraging localities was greater than in the "natural" burn regimes.  In particular, foraged localities had remarkably uniform distributions of successional stage.  The non-foraging localities would show peaks at different successional stages -- depending on the age of the last burn -- but the frequency of whatever stage was present would typically fall off from whatever the characteristic mean was.  In contrast, the successional stage distribution we observed in the foraged localities looked like a mixture of several different smaller "natural" localities.  This is because that's essentially what they are!

When people burn for subsistence needs, they tend to start lots of small fires at many different times.  What results is a mosaic of successional stages. But this is really a local re-arrangement.  When we looked at the landscape level, the diversity of habitats looks very much like that of the smaller foraging localities.

One interesting implication of this work is that the human impact on the Australian environment -- grassland engineering, if you like -- was likely to only be substantial following the establishment of more intensified aboriginal economies (approximately 1500 years ago).

It may seem counter-intuitive that hunting with fire promotes biodiversity, but that really seems to be what is happening here.  The results have clear conservation and management implications. Excessive fire suppression in highly flammable grasslands is probably not a good idea.  Now we need to measure the actual species diversity rather than simply the diversity of habitat types.

See Rebecca's fire ecology website for more on this research and reprints of various papers.  The full text of our paper can be found here.