To understand how ectotherms will respond to warming temperatures, we require information on thermal habitat quality at spatial resolutions and extents relevant to the organism. Measuring thermal habitat quality is either limited to small spatial extents, such as with ground‐based 3D operative temperature (Te) replicas, representing the temperature of the animal at equilibrium with its environment, or is based on microclimate derived from physical models that use land cover variables and downscale coarse climate data. We draw on aspects of both these approaches and test the ability of unoccupied aerial vehicle (UAV) data (optical RGB) to predict fine‐scale heterogeneity in sub‐canopy lizard (Anolis bicaorum) Te in tropical forest using random forest models. Anolis bicaorum is an endemic, critically endangered, species, facing significant threats of habitat loss and degradation, and work was conducted as part of a larger project. Our findings indicate that a model incorporating solely air temperature, measured at the centre of the 20 × 20 m plot, and ground‐based leaf area index (LAI) measurements, measured at directly above the 3D replica, predicted Te well. However, a model with air temperature and UAV‐derived canopy metrics performed slightly better with the added advantage of enabling the mapping of Te with continuous spatial extent at high spatial resolutions, across the whole of the UAV orthomosaic, allowing us to capture and map Te across the whole of the survey plot, rather than purely at 3D replica locations. Our work provides a feasible workflow to map sub‐canopy lizard Te in tropical environments at spatial scales relevant to the organism, and across continuous areas. This can be applied to other species and can represent species within the same community that have evolved a similar thermal niche. Such methods will be imperative in risk modelling of such species to anthropogenic land cover and climate change.

中文翻译：

---

无人飞行器作为绘制热带环境中蜥蜴工作温度的工具

为了了解变温动物如何响应变暖的温度，我们需要有关生物体相关空间分辨率和范围的热栖息地质量信息。测量热栖息地质量要么仅限于较小的空间范围，例如地面 3D 操作温度（时间e）复制品，代表动物与其环境平衡的温度，或者基于使用土地覆盖变量和小尺度粗略气候数据的物理模型得出的小气候。我们利用这两种方法的各个方面，并测试无人飞行器（UAV）数据（光学 RGB）预测亚冠蜥蜴精细尺度异质性的能力（双高安乐蜥）时间e在热带森林中使用随机森林模型。双高安乐蜥是一种特有的、极度濒危的物种，面临着栖息地丧失和退化的重大威胁，这项工作是作为一个更大项目的一部分进行的。我们的研究结果表明，仅包含在 20 × 20 m 图中心测量的气温和在 3D 复制品正上方测量的地面叶面积指数 (LAI) 测量值的模型预测时间e出色地。然而，具有气温和无人机衍生的冠层指标的模型表现稍好，并且具有能够绘制时间e具有高空间分辨率的连续空间范围，跨越整个无人机正射马赛克，使我们能够捕获和绘制地图时间e跨越整个调查地块，而不是纯粹在 3D 复制位置。我们的工作提供了绘制亚冠蜥蜴地图的可行工作流程时间e在热带环境中，在与生物体相关的空间尺度上，以及跨越连续的区域。这可以应用于其他物种，并且可以代表同一群落内进化出类似热生态位的物种。在对此类物种对人为土地覆盖和气候变化进行风险建模时，此类方法至关重要。