How to study a predator that only eats a few meals a year: high-frequency accelerometry to quantify feeding behaviours of rattlesnakes (Crotalus spp.)
How to study a predator that only eats a few meals a year: high-frequency accelerometry to quantify feeding behaviours of rattlesnakes (Crotalus spp.)
复制标题
DOI：
10.1186/s40317-023-00332-3
复制
发表时间：
2023-05-13
影响因子：
2.7
通讯作者：
Clark, Rulon W.
中科院分区：
文献类型：
--
作者：
Hanscom, Ryan J.;DeSantis, Dominic L.;Clark, Rulon W.
背景许多蛇是低能量的捕食者，它们利用隐蔽来伏击猎物。这些物种中的大多数很少进食，对人类等大型脊椎动物的存在很敏感，并且一生中的大部分时间都在隐藏中度过。这使得直接观察摄食行为具有挑战性，以前为记录自由放养的蛇的捕食行为而开发的方法具有严重的局限性。生态学家越来越多地使用动物传播的加速度计来量化自由放养动物的活动和瞬间行为，但它们在蛇中的应用仅限于记录基本的行为状态（例如，活动与非活动）。高频加速度计可以提供新的洞察这一重要群体的捕食者的行为，在这里，我们提出了一种新的方法来量化的三种蝰蛇（响尾蛇属）的摄食行为的关键方面。和评估的可转移性的分类模型在这些species.ResultsWe使用开源软件创建特定物种的模型，分类运动，静止，捕食性打击，和猎物吞咽具有高精度，准确性和召回。此外，我们确定了一种低成本，可靠，非侵入性的附着方法，用于将加速度测量装置放置在蛇的前部，这可能是准确分类这些物种中不同行为所必需的。然而，物种特异性模型在我们的跨物种comparation.ConclusionsOverall，我们的研究表明了强大的潜力，使用加速度计记录关键的喂养行为的蛇，难以直接观察。此外，我们提供了一个“端到端”的模板，用于识别使用高频加速度计的蝰蛇的觅食生态中涉及的重要行为。我们强调了一种方法的附件的加速度计，一种技术来模拟饲养事件在圈养，和模型选择过程中使用生物相关的窗口大小在一个开放访问的软件分析加速度数据（AcceleRater）。虽然我们无法获得跨物种的通用模型，但如果从不同体型和不同背景的蛇中纳入更多数据（即，通过自然栖息地移动），可以开发具有更高可转移性的通用模型。
BackgroundMany snakes are low-energy predators that use crypsis to ambush their prey. Most of these species feed very infrequently, are sensitive to the presence of larger vertebrates, such as humans, and spend large portions of their lifetime hidden. This makes direct observation of feeding behaviour challenging, and previous methodologies developed for documenting predation behaviours of free-ranging snakes have critical limitations. Animal-borne accelerometers have been increasingly used by ecologists to quantify activity and moment-to-moment behaviour of free ranging animals, but their application in snakes has been limited to documenting basic behavioural states (e.g., active vs. non-active). High-frequency accelerometry can provide new insight into the behaviour of this important group of predators, and here we propose a new method to quantify key aspects of the feeding behaviour of three species of viperid snakes (Crotalus spp.) and assess the transferability of classification models across those species.ResultsWe used open-source software to create species-specific models that classified locomotion, stillness, predatory striking, and prey swallowing with high precision, accuracy, and recall. In addition, we identified a low cost, reliable, non-invasive attachment method for accelerometry devices to be placed anteriorly on snakes, as is likely necessary for accurately classifying distinct behaviours in these species. However, species-specific models had low transferability in our cross-species comparison.ConclusionsOverall, our study demonstrates the strong potential for using accelerometry to document critical feeding behaviours in snakes that are difficult to observe directly. Furthermore, we provide an 'end-to-end' template for identifying important behaviours involved in the foraging ecology of viperids using high-frequency accelerometry. We highlight a method of attachment of accelerometers, a technique to simulate feeding events in captivity, and a model selection procedure using biologically relevant window sizes in an open-access software for analyzing acceleration data (AcceleRater). Although we were unable to obtain a generalized model across species, if more data are incorporated from snakes across different body sizes and different contexts (i.e., moving through natural habitat), general models could potentially be developed that have higher transferability.