Comparing Methods of Detecting an Elusive Dasyurid Marsupial, the Threatened Julia Creek Dunnart (<i>Sminthopsis douglasi</i>), in Central Western Queensland, Australia

The Julia Creek dunnart, <i>Sminthopsis douglasi</i>, is a small, threatened carnivorous marsupial occurring in scattered populations in the grasslands of central and northwestern Queensland, Australia. The distribution of the species is largely unknown due to sporadic survey efforts and its historically low detection using traditional live trapping methods. There is an urgent need to determine the best methods of detection to optimise survey methodologies and more effectively manage species conservation efforts. In this study, we compared the effectiveness of live (Elliott) traps, baited white flash camera traps and thermal imagery binocular surveying for detecting <i>S. douglasi</i>. We deployed 40 white flash camera traps at two sites in Bladensburg National Park (south of Winton), where the species is known to occur, for three consecutive periods between June and November 2022. Four comparative sessions of live trapping were undertaken between April and August 2022 at the same locations. During the live trapping periods, a total of 12 nights of surveying were conducted with thermal imagery binoculars in a preliminary assessment of the technique. The total live trapping effort was 3600 trap nights (approximately 700 trap nights per site in each trapping event). Live trapping resulted in 12 detections of individual <i>S. douglasi</i> from 19 total captures. The highest trap success on a given trapping session was 1.71%, and overall trap success from both sites across all sessions was 0.53%. In comparison, baited camera traps (deployed facing the ground at 70 cm range) took 1,269,884 images over 5383 trap nights. There were 11 confirmed images of <i>S. douglasi</i>, on three individual occasions, which represented 2.10% of all small mammal captures and just 0.0009% of the total images. Four species of small mammals were detected using camera traps, whereas live trapping detected only two species. No small mammals were detected on any of the 12 thermal binocular surveys. Overall, our study highlights the comparative high utility of traditional live trapping for detecting <i>S. douglasi</i>. This research provides a framework for ongoing monitoring of the Bladensburg National Park population. It will be more broadly beneficial for informing the best detection techniques of <i>S. douglasi</i> in ongoing work investigating the overall distribution of the species. Similar studies assessing multiple detection methods for small terrestrial mammals have shown an advantage of white flash camera traps compared to other traditional detection techniques. Our contrasting results serve as a reminder that the utility of different techniques for detecting small mammals is best assessed on a species-by-species basis.