Introduction
Curbing worldwide wildlife population declines will necessitate the protection of their habitat (Tilman et al. 2017). Aside from political willingness, targeted habitat protection requires robust baseline information about wildlife that use and occupy it. Collating such information, however, remains a challenging and costly endeavour. To tackle these problems, scientists have been working hard developing innovative technologies that improve the cost-effectiveness and scalability of data collection on wildlife assemblages, including employing thermal drones, detection dogs, radars, camera traps and bioacoustics devices (Cristescu et al. 2015, Hagens et al. 2018, Hüppop et al. 2019, Beaver et al. 2020, Law et al. 2020). Whilst these technologies have greatly expanded our capabilities to collect more accurate baseline information on wildlife presence, the collation of data on wildlife assemblages still comes at a significant technological and human cost. To seek to identify the presence of multiple wildlife species, ecologists must deploy a multitude of methods in synchrony, which can include bioacoustics arrays, baited traits, camera traps and human-led fauna transects (Mena et al. 2021). Hence, affordable and scalable technology for the detection and monitoring of wildlife assemblages across the landscape is a goal we should seek to achieve as it would improve our ability to detect and track changes in wildlife biodiversity patterns (e.g. species richness and community composition), identify wildlife species or habitat of high conservation priority, and underpin the development of successful management strategies.
Environmental DNA, known as eDNA is a promising technology to address this issue.  eDNA refers to the DNA that is shed into the environment as wildlife move throughout the landscape, leaving previously unbeknownst traces of their presence (Thomsen and Willerslev 2015). Already heavily deployed in aquatic systems for at least two decades, the practicability of utilising eDNA to detect species presence and assemblages is clearly established (Hinz et al. 2022). For example, in 2016, aquatic eDNA technology was first employed as part of environmental impact assessments, demonstrating its capacity to meet regulatory standards and obligations outlined by the International Association for Impact Assessment (Hinz et al. 2022). However, to date, there is no such established terrestrial analogue although recent studies have provided promising results. Terrestrial wildlife’s eDNA has, for instance, been detected from samples collected from spider webs, permafrost, blood, snow, soil, honey and aerosol spray runoff (Andersen et al. 2012, Folloni et al. 2012, Schnell et al. 2012, Valentin et al. 2018, Ribani et al. 2020, Gregorič et al. 2022). Two most recent studies have also demonstrated that airborne eDNA can be used to reassemble zoological communities (Clare et al. 2022, Lynggaard et al. 2022). Together, evidence suggest that airborne eDNA could be a promising avenue for the identification of terrestrial wildlife assemblages as it is i) non-invasive, ii) scalable, and iii) comparatively cost-effective given it holds the potential to target species assemblages compared to single species targeted approaches. The feasibility of using airborne eDNA to identify wildlife assemblages under natural conditions remains yet to be tested.
Here, we sought to test the applicability of airborne eDNA particles for the detection of an endangered Australian species, the koala (Phascolarctos cinereus ), and its co-occurring terrestrial mammalian community in a natural setting. We demonstrate its successful application to detecting koalas as well as other mammalian species in a natural setting and discuss future steps for its continued improvement and optimisation.