Data from: Evaluation of an acoustic telemetry transmitter designed to identify predation events

The field of acoustic telemetry has evolved rapidly and now permits the remote sensing of animal behaviour, movement, physiology and survival in environments and species not previously possible. However, an inability to detect when a telemetered animal is consumed by a predator can complicate accurate interpretation of telemetry data. Here, we describe efforts to test two generations of a novel prototype acoustic telemetry transmitter designed specifically to detect predation.
Testing involved either staged predation events where tagged prey (Rainbow Trout Oncorhynchus mykiss and Yellow Perch Perca flavescens) were fed to captive Largemouth Bass Micropterus salmoides, or false positive testing where prey fish were tagged and held without risk of predation. Metrics of interest were (a) the rate of correctly identifying predation events, (b) signal lag (i.e. the time required to detect a predation event), (c) tag retention time in the predator's gut, and (d) the rate of false positive triggering in both live and dead prey fishes.
Staged predation events were successfully identified in 61/65 and 52/55 trials for generation 1 and 2 tags, respectively. Signal lag time was reduced in generation 1 tags (generally between 1 - 9 hours) relative to generation 2 (3 - 29 hours); although signal lag was highly variable. A generalized linear mixed model indicated strong evidence that signal lag and tag retention were both negatively correlated with water temperature, but were not affected by prey species and only slightly by individual predator traits. There was preliminary evidence that prey size may be an important determinant of both signal lag and tag retention. False positives in live fish were absent after 120 days for generation 1 tags (n=31), however rates were significantly higher (10/44) after only 66 days for generation 2 tags. False positives in dead fish suggested that 20% of generation 2 predation tags would falsely trigger 2-3 days post-mortem.
Testing of the novel predation tags was encouraging however further testing is recommended. Predation tags will be an important contribution to the field of acoustic telemetry; permitting improved data interpretation and less subjective estimates of predation rates in biotelemetry studies.

声学遥测(acoustic telemetry)领域发展迅猛，如今已可在此前难以实现研究的环境与物种中，远程监测动物的行为、活动、生理状态与存活情况。然而，当前技术无法检测被遥测标记的动物被捕食者捕食的时刻，这会给遥测数据的精准解读带来复杂挑战。本研究介绍了针对捕食事件检测专门设计的两代新型声学遥测发射机原型的测试工作。

试验分为两类：一是模拟捕食事件，即将标记好的猎物（虹鳟*Oncorhynchus mykiss*、黄鲈*Perca flavescens*）投喂给圈养的大口黑鲈*Micropterus salmoides*；二是假阳性测试，即对猎物鱼进行标记并饲养，使其无被捕食风险。本研究关注的评估指标包括：(a) 捕食事件的正确识别率；(b) 信号延迟时长（即检测到捕食事件所需的时间）；(c) 标记物在捕食者消化道内的留存时间；(d) 活体与死体猎物鱼的假阳性触发率。

对于第1代与第2代标记物，模拟捕食事件的正确识别率分别为61/65与52/55。第1代标记物的信号延迟时长通常介于1~9小时，短于第2代的3~29小时，但信号延迟的波动性极强。

广义线性混合模型(generalized linear mixed model)分析结果显示，信号延迟与标记物留存时间均与水温呈显著负相关，且不受猎物物种影响，仅受捕食者个体特征的轻微影响。另有初步证据表明，猎物体型可能是影响信号延迟与标记物留存时间的重要因素。

第1代标记物在活体鱼中的假阳性触发情况在120天后完全消失（n=31），而第2代标记物仅在66天后的假阳性触发率就显著升高（10/44）。死体鱼的假阳性测试结果显示，20%的第2代捕食标记物会在动物死后2~3天触发假阳性信号。

新型捕食标记物的试验结果令人鼓舞，但仍建议开展进一步试验。捕食标记物将为声学遥测领域作出重要贡献，可提升生物遥测(biotelemetry)研究中的数据解读精度，并减少捕食率估算的主观性。