Context. Aerial distance sampling surveys are a widely used approach for estimating the abundance of wildlife populations, including kangaroos and other macropods. When analysing such data, design-based estimates of abundance extrapolate density from the sampled region to the broader area. In contrast, model-based approaches that use relationships between population density and habitat variables can deliver greater precision and ecological insight into population estimates.

Aims. This study aimed to evaluate recent spatio-temporal trends in the abundance of grey kangaroos (Macropus giganteus and M. fuliginosus) in Victoria, Australia using both design-based and model-based distance sampling.

Methods. Aerial surveys were conducted in seven management zones spread across the non-forested parts of Victoria on four occasions between 2017 and 2022. Distance sampling data were analysed to obtain design-based and model-based estimates of population size and relative precision across the entire study area, and for each management zone for each of the four surveys.

Key results. Model-based and design-based estimates of abundance proved to be broadly comparable, but with mostly higher precision for the model-based estimates at both statewide and management zone scales. The model-based approach revealed relationships between transect-level abundance and land-use type, normalised difference vegetation index (NDVI), annual rainfall and the amount of ecotonal habitat in the vicinity of the transects. Between 2017 and 2022 (a period coinciding with a prolonged La Niña event), the total abundance of grey kangaroos increased by approximately 87% (based on the model-based inferences) despite an ongoing harvesting program.

Conclusions. In addition to improved precision, model based estimation revealed relationships between habitat variables and transect-level abundance, allowing prediction of abundance at small scales.

Implications. The improved precision of model-based estimates of abundance will provide greater confidence in kangaroo management decisions such as setting of statewide harvest quotas. Model-based estimates at smaller spatial scales will similarly support localised management decisions.