Aborigine-managed forest, savanna and grassland: Biome switching in montane eastern Australia

Aim: To assess hypotheses about the role of anthropogenic fire in the maintenance and origin of a fine-scale vegetation mosaic of rain forest, eucalypt savanna and grassland. Location: Bunya Mountains, subtropical eastern Australia. Methods: A time series of vegetation maps was compiled from historical and recent aerial photography and field surveys. Geospatial models were constructed of environmental domains for rain forest, savanna and grassland, and for areas of biome change. Grassland soils were analysed for carbon isotope ratios (δ¹³C), and radiocarbon (¹⁴C) dates were acquired for bulk samples from a range of depths. Results: Analysis revealed weak associations between topography and the distribution of rain forest, savanna and grassland, and their patterns of recent change. Grassland occupied an environmental domain intermediate between rain forest and savanna and was more than four times as likely to occur within a matrix of rain forest rather than savanna. There was a large proportional reduction in the area of both grassland (-35%) and savanna (-19%) between 1961 and 2006 because of the expansion of rain forest. However, the greater initial extent of savanna meant that the areal loss of savanna was an order of magnitude greater than for grassland (1433 vs. 146 ha). There was no evidence of abrupt changes in δ¹³C in grassland soil profiles, indicating stability of the vegetation over the last 2000 years. Main conclusions: There is no simple gradient in 'tree suitability' from rain forest, through savanna, to treeless grassland on the Bunya Mountains. A general absence of fire since the 19th century has greatly reduced the extent of grassy savanna and grassland formations, to the advantage of rain forest. These results support the hypothesis that the vegetation mosaic on the Bunya Mountains is a cultural artefact and testament to millennia of skilful and persistent burning. We could not conclusively reject the hypothesis that the grasslands are Pleistocene relicts that have declined throughout the Holocene; nonetheless, an explanation more consistent with the evidence overall is that the grasslands must have had periods of expansion during the Holocene, probably as a consequence of severe fires that have destroyed patches of rain forest.