Ecological implications of standard fire-mapping approaches for fire management of the World Heritage Area, Fraser Island, Australia

Abstract

The characterisation of spatiotemporal fire patchiness is requisite for informing biodiversity conservation management in many landscape settings. Often, conservation managers are reliant on manually derived fire-history mapping products that delineate fire perimeters. An alternative standard approach concerns the application of remote sensing, typically using band combination indices obtained from relatively fine-scale imagery sensors. For Fraser Island, a World Heritage Area in subtropical, fire-prone eastern Australia, we contrast diagnostic fire-regime characteristics for different vegetation types over a 20-year period (1989–2008) as derived from historical manual, and remotely sensed, fire-mapping approaches. For the remote sensing component we adapt a commonly used approach utilising a differenced normalised burn ratio (dNBR) index derived from Landsat Thematic Mapper imagery. Manual mapping resulted in overestimation of fire-affected area (especially large fires) and fire frequency, whereas the dNBR procedure resulted in underestimation of fire-affected area under low fire-severity conditions, and overestimation of fire patchiness. Of significance for conservation management, (1) age class and related distributions for flammable vegetation types differed markedly between the two mapping approaches, (2) regardless, both methods demonstrated that substantial fuel loads had accumulated in flammable vegetation types by the end of the study period and (3) fuel age was shown to have a more significant effect than did seasonality on the incidence of very large (>1000 ha) fires. The study serves as an introduction to ongoing research concerning the measurement and application of fire patchiness to conservation management in flammable eastern Australian vegetation types.