Species distribution models for conservation planning in fire‐prone landscapes

Abstract

Species distribution models are an essential tool for biodiversity conservation, with important applications such as spatial prioritisation of conservation actions and elucidating relationships between environmental predictors and species responses. These models are most useful to conservation managers when they include factors that can be readily manipulated, such as fire. In this study, we collated a comprehensive dataset of mammal records from a fire-prone region in south-east Australia where mammals have suffered declines in recent decades. We used species distribution modelling to (1) determine the relative influence of climate, fire, vegetation and topography on ground-dwelling mammal distributions; (2) determine species responses to time since fire, and; (3) provide spatial predictions of habitat suitability for conservation planning. Climate was the predominant driver of habitat suitability for most species, although other factors were influential in some cases. Time since fire was an important factor driving the distribution of only two of 16 modelled species which were more likely to be recorded in recently burnt vegetation. Habitat suitability varied spatially among species however multi-species habitat suitability was highest in the drier and hotter eastern section of the landscape, highlighting a key area for conservation efforts. The outputs from our models can be used for practical conservation actions such as finding new populations or identifying sites for reintroductions. We conclude that presence-only species distribution models are useful for determining species fire responses, complementing more systematic methods, and that including dynamic variables, such as time since fire, can increase their conservation relevance.