Factors confounding koala habitat mapping at multiple decision‐making scales

Abstract

SummaryThe Koala (Phascolarctos cinereus) is an arboreal marsupial found throughout southeastern Australia. A high risk of extinction in some areas requires adherence to legislation mandating protection of high‐quality habitat. The Koala primarily eats leaves of the Eucalyptus genus and wildlife managers are, in most cases, currently dependent on habitat maps depicting food resource availability derived from low‐resolution regional‐extent (~100,000 hectares) vegetation maps. These maps might not capture resource variability at planning extents (0.1–300 ha), or at resolutions where Koala ecology processes operate (5–50 ha). Consequently, potentially high‐quality habitat may not be considered in planning decisions. Whether low‐resolution habitat mapping adequately captures food resource variability is therefore relevant to map users at non‐regional extents. We examined four different broad‐scale low‐resolution habitat maps within a small area (150 ha) in southeast Queensland and compared the classifications of each. Map units were generally ranked in the same hierarchical order; however, differences in methodologies meant that, between maps, some units varied widely in assigned habitat quality. Two maps were assessed using quantitative Queensland Herbarium data, and field data we collected. Assessment using overstorey species emphasised habitat quality differences between habitat map units better than the alternative whole‐of‐canopy method. Maps relying on interpretation of vegetation descriptions for habitat class definition sometimes overestimated food tree cover proportions and stakeholders need to be cognisant of these limitations. One method assigned Very Low Suitability to a 60‐ha focal study site containing evidence of Koala use, and we strongly recommend that field validation should be an integral part of habitat management at these scales. We conclude that, rather than mapping approaches encompassing several habitat attributes within a single map, layered datasets, each with specific attributes, would provide greater utility for stakeholders by allowing them to use layers individually or in combination as required.