Thermal imaging outshines spotlighting for detecting cryptic, nocturnal mammals in tropical rainforests

Abstract

Context Thermal imaging has been shown to be a valuable technique for detecting a range of terrestrial mammals across different environments. The limited studies looking at its effectiveness in detecting arboreal mammals in rainforest ecosystems have had mixed success due to the complexity of the environment and limitations of the technology itself. Aims We investigated whether using a hand-held thermal imaging device would detect more individuals of six species of nocturnal arboreal mammal in tropical rainforests than the most-used detection method of spotlighting. We determined whether environmental variables effecting either equipment operation or mammal behaviour would influence these results. Methods We surveyed eight transects across the Wet Tropics of northern Queensland for six species of arboreal mammals using both a hand-held thermal imager and a spotlight. We used a measure of underestimation to compare counts of individual species, and then modelled total mammal counts with detection method and environmental variables to find the best approximating model. Key results Spotlighting underestimated the total number of each species by between 33 and 100% when compared with thermal imaging. Detection method alone without any environmental interaction term provided the best approximating model (AICc = 275.58, marginal pseudo R2 = 0.286), with thermal imaging technology detecting almost double the number of our target individuals (12.3 ± 1.76) compared with spotlighting (6.7 ± 1.02). Conclusions Despite recorded operational limitations, thermal imaging technology greatly improved our ability to locate both small and large nocturnal, arboreal mammals, including a species that is rarely observed in the wild. Implications The potential to not only improve detection of nocturnal, arboreal mammals but also improve cryptic species distribution and abundance measures suggests thermal imaging technology is an important tool for use globally across rainforests environments.