Context: The development of reliable and cost-efficient survey techniques is key to the monitoring of all wildlife. One group of species that presents particular challenges for monitoring is the arboreal mammals. Traditional techniques for detecting these species often yield low detection probabilities (detectability) and are time-consuming, suggesting the potential for novel methods to enhance our understanding of their distribution, abundance and population trajectories. One technique that has been shown to increase detectability in a range of terrestrial species is thermal imaging, although it has rarely been applied to arboreal species. The true conservation status of Lumholtz’s tree-kangaroo (Dendrolagus lumholtzi) is uncertain because of low detectability under typical survey techniques, and a more suitable method is required to enable effective monitoring of the species, making it an ideal candidate for the present study.

Aims: We aimed to compare the success and cost-effectiveness of surveys utilising thermal imaging with two traditional methods, namely, spotlighting and daytime surveys, so as to optimise monitoring of D. lumholtzi.

Methods: We conducted surveys at 10 sites in Queensland (Australia) where D. lumholtzi was known to occur, by using each method, and modelled both the detectability of D. lumholtzi and the cost-effectiveness of each technique.

Key results: Detectability of D. lumholtzi was significantly higher with the use of thermal imaging than it was with the other survey methods, and thermal detection is more cost-effective. In average survey conditions with a trained observer, the single-visit estimated detectability of D. lumholtzi was 0.28 [0.04, 0.79] in a transect through rainforest, by using thermal imaging. Using only spotlights, the detection probability was 0.03 [0, 0.28] under the same conditions.

Conclusions: These results show that incorporating thermal technology into monitoring surveys will greatly increase detection probability for D. lumholtzi, a cryptic arboreal mammal.

Implications: Our study highlighted the potential utility of thermal detection in monitoring difficult-to-detect species in complex habitats, including species that exist mainly in dense forest canopy.

背景：开发可靠且具有成本效益的调查技术是监测所有野生动物的关键。对监测提出特别挑战的一组物种是树栖哺乳动物。用于检测这些物种的传统技术通常会产生低检测概率（可检测性）并且非常耗时，这表明新方法具有增强我们对其分布、丰度和种群轨迹的理解的潜力。一种已被证明可以提高一系列陆地物种可检测性的技术是热成像，尽管它很少应用于树栖物种。鲁姆霍兹树袋鼠（Dendrolagus lumholtzi）的真实保护现状) 是不确定的，因为在典型的调查技术下可检测性低，需要更合适的方法来有效监测物种，使其成为本研究的理想候选者。

目的：我们旨在比较利用热成像的调查与两种传统方法（即聚光和白天调查）的成功率和成本效益，以优化对D. lumholtzi 的监测。

方法：我们使用每种方法在昆士兰（澳大利亚）已知存在D. lumholtzi 的10 个地点进行了调查，并对D. lumholtzi的可检测性和每种技术的成本效益进行了建模。

主要结果：使用热成像技术对D. lumholtzi的探测能力明显高于其他调查方法，并且热探测更具成本效益。在由训练有素的观察员进行的平均调查条件下，通过使用热成像，在穿过雨林的横断面中，D. lumholtzi的单次访问估计可探测性为 0.28 [0.04, 0.79]。仅使用聚光灯，相同条件下的检测概率为0.03 [0, 0.28]。

结论：这些结果表明，将热技术纳入监测调查将大大增加D. lumholtzi（一种神秘的树栖哺乳动物）的检测概率。

意义：我们的研究强调了热探测在监测复杂栖息地中难以探测的物种方面的潜在效用，包括主要存在于茂密森林冠层中的物种。