Modeling variability in the fire response of an endangered bird to improve fire-management.

Conservation managers regularly burn vegetation to regenerate habitat for fire-dependent species. When determining the time since fire at which to burn, managers model change in a species' occurrence over time, post-fire (fire-response curve) and identify the time since fire associated with decline in occurrence. However, where species exhibit variability in their fire response across space, using a single fire-response curve to determine the timing of burns may lead to burning habitat at an inappropriate time since fire. We tested if elevation, local topography, soil properties, vegetation type or evapotranspiration affect the fire response of the endangered Mallee Emu-wren Stipiturus mallee and its hummock-grass habitat Triodia scariosa in southeastern Australia (n = 217). Previous work on the Mallee Emu-wren found a unimodal fire response with decline in occurrence at ~30-50 yr since fire and a time window of occurrence of ~30 yr. We found that time since fire and elevation interact to affect the Mallee Emu-wren fire response. At high elevations (55-98 m), Mallee Emu-wrens declined in occurrence at ~50 yr since fire, with a time window of occurrence of 20-40 yr. However, at low elevations (28-55 m), Mallee Emu-wrens showed no decline in occurrence with increasing time since fire with a time window of occurrence of up to 107 yr. Extent cover of Tall T. scariosa showed similar patterns to the Mallee Emu-wren, indicating that vegetation structure is a likely driver of variability in the Mallee Emu-wren fire response. We speculate that the effect of low elevation is mediated by increased soil nutrient and water availability for key plants. We used our findings to map the appropriate time since fire at which to burn to regenerate habitat for the Mallee Emu-wren across the study region. We recommend no burning for regeneration across one-third of potential habitat, because the Mallee Emu-wren showed no decline in occurrence in these areas. We recommend managers model variability in species' fire responses across space to improve the timing of burns for regeneration.