Quantifying the direct fire threat to a critically endangered arboreal marsupial using biophysical, mechanistic modelling

Abstract

AbstractHistorically unprecedented areas of forest habitat have been impacted by fire, as climate change and other anthropogenic disturbances drive increases in fire burned area and severity. Although 88% of Australia's [threatened] land mammals are threatened by inappropriate fire regimes, calculations of animal mortality resulting from specific events have been impeded by knowledge gaps relating to both the direct (first‐order) and long‐term (second‐order) effects of fire on different species. This study addresses the need for a quantified, mechanistic understanding of first‐order effects, presenting an extension of the Fire Research and Modelling Environment (FRaME) to allow prediction of species‐specific mortality. FRaME is demonstrated and tested here by replicating an incident in which a prescribed burn caused 77% mortality of a population of the critically endangered ngwayir (Pseudocheirus occidentalis, Pseudocheiridae). FRaME correctly predicted heavy mortality (62–79%) arising from partial and full‐thickness burns and asphyxiation due to burns in the respiratory tract. Mortality varied with animal fire‐avoidance strategies (p < 0.001) and the thickness of tree hollow walls (r = −0.95, p < 0.001). Although management guidelines specified low intensity fire, mortality had no significant relationship with Byram intensity and larger flames due to ‘torching’ were most frequent when fire spread was slowest. FRaME modelling predicted that individuals would be impacted by temperatures exceeding 500°C for several minutes. Fire management that is premised on discredited notions of fire behaviour and overly simple models can lead to catastrophic management outcomes such as those documented here. FRaME addresses this need by providing a platform to account for heterogeneous fire behaviour as well as animal behaviour and habitat quality, calculating fire risk to fauna and guiding management that maximizes safe habitat.