Post-fire live and dead fuel flammability stabilises Eucalyptus forest-sedgeland boundaries in southern Tasmania

The mosaics of forest and treeless vegetation in western and southern Tasmania have been explained by the ‘ecological drift’ model, an exemplar of terrestrial Alternative Stable State (ASS) theory. This theory posits that vegetation patterns are controlled by fire, and that anomalous changes to fire frequency can cause rapid change in landscape-scale vegetation patterns. We used a variety of methods to test key predictions of the ecological drift model related to landscape flammability, namely that treeless sedgelands are more flammable than adjacent Eucalyptus forest, and that fires cause all vegetation types to become more flammable. A LiDAR survey before and after an extensive wildfire in 2019 revealed loss of near-surface fuels, but vegetation structure was maintained across Eucalyptus forest, adjacent sedgeland and the scrub ecotone communities. Field sampling showed that fuel types and fuel loads differed significantly amongst the three communities, with the highest loads of surface and near surface fuels in the forest and the least in the sedgeland. These fuel loads were reduced by the fire, to near zero in sedgeland, indicating landscape flammability was substantially reduced for at least five years post-fire in sedgeland. While laboratory testing found only small differences in litter flammability among the communities, live foliage of the dominant species was most flammable in sedgeland. Similarly, sedgeland had the driest microclimate post-fire, with three times as many days when litter would be dry enough to burn as forest. Our synthesis of these findings suggests that counter to predictions of the ecological drift model, short-term fire risk in sedgeland is reduced for at least five years because of greatly reduced fuel loads. During this period, sedgeland is likely to be less flammable than scrub or forest, with its lack of fuel outweighing its drier post-fire microclimate and more flammable dominant species. Consequently, we concluded that a ‘stable fire cycle’ model based primarily on edaphic differences, with fuel flammability a sharpening switch, is more apt to explain forest -sedgeland boundaries than the ASS based ecological drift model.