ATBC Online Web Program

O13-2
Patterns of fire and biodiversity in Australian tropical savanna:  A critical analysis of the relationships between satellite derived fire histories and terrestrial fauna

Tuesday, 25 June 2013: 10:35
La Paz - B East (Herradura San Jose)
Justin James Perry , Ecosystems Sciences, CSIRO, Australia
Eric Vanderduys , Ecosystems Sciences, CSIRO, Australia
Genevieve Perkins , Ecosystems Sciences, CSIRO, Australia
Anders Zimny , Ecosystems Sciences, CSIRO, Australia
There is an increasing trend in ecological studies to use remotely sensed data to describe biological systems and landscape health.  Often the accuracy of these data and the links to ‘real’ natural systems are tenuous and untested.  In northern Australia there are clear negative relationships between high fire frequency and terrestrial fauna richness and abundance in savanna ecosystems.  Theoretically, the published relationships should transpose onto broad scale satellite derived fire histories which could act as proxy for biodiversity health.  We compare three years of systematically collected terrestrial biodiversity survey data from Cape York Peninsula, northern Australia (mammals, reptiles, frogs and birds – floristics and structure) with 12 years of MODIS derived fire scars (fire frequency, extent and time since fire).  Terrestrial fauna sites were selected to capture the variance of MODIS derived fire frequency within a variety of common broad vegetation groups on Cape York Peninsula. The variance in richness and abundance of fauna between sites of the same habitat was not strongly influenced by the satellite derived fire histories but by micro-habitat at the sites.  This is particularly true for species with small home ranges where fine scale structural heterogeneity drives population dynamics. Fire frequency was important at a landscape scale (between habitat types) but generally reflected the relative fire proneness of each habitat.  Low fire frequency sites in high fire frequency areas were often atypical of the vegetation type, i.e. fire suppressed by intense grazing. MODIS derived fire histories are used frequently by land managers to plan prescribed fires for biodiversity conservation and are often used as a proxy for biodiversity health.  We propose the resolution of the commonly used imagery (MODIS) is too coarse spatially (250m) and too short temporally (10 years) to adequately describe the impact of fire on fauna in northern Australia.