Data from: Digital recorders increase detection of Eleutherodactylus frogs

Improved methods for large-scale and long-term amphibian monitoring projects would aid species status assessments and identification of potentially declining species (e.g., Stuart et al. 2004). Acoustical surveys allow researchers to quickly determine species presence and calling activity level (Lips et al. 2001; Zimmerman 1994). The North American Amphibian Monitoring Program, a large-scale monitoring effort, uses acoustical surveys to monitor species in ponds and wetlands (United States Geological Survey 2004). A challenge of this approach is accurate species identification, which is particularly problematic in a tropical area with high anuran species diversity (e.g., anuran species richness in the Neotropics was greater than in the Neartic, Duellman 1999). Furthermore, abundant or loud species may lower the detection probability of other species. An assumption that needs to be tested in acoustical surveys is that detection is highly correlated with the species presence or activity. Without an appropriate correction for detection probability, data will not accurately reflect the status of the species (MacKenzie et al. 2002). This may result in inappropriate management decisions that could waste time and money. Digital recorders can be used as a standard method for species detection and determination of calling activity level in amphibian choruses in long-term monitoring projects. Digital recordings can be stored and transferred easily and can allow experts to analyze the recordings at their convenience. In addition, loud or abundant species can be filtered using computer software to increase the detection of other species. I tested the hypothesis that detection of species presence and calling activity level in anuran choruses will be higher when recordings are analyzed with the help of computer software compared to only hearing the recordings. The resulting increase in detection should result in improved population status data and reduced false negatives during species inventories (i.e., when a species is not detected due to interference).

适用于大规模长期两栖动物监测项目的优化方法，可为物种现状评估以及潜在濒危物种的识别工作提供助力（如Stuart等人2004年的研究）。声学调查可帮助研究人员快速确定物种存在情况及其鸣叫活动强度（Lips等人2001年；Zimmerman 1994年）。北美两栖动物监测计划（North American Amphibian Monitoring Program）作为一项大型监测项目，采用声学调查手段对池塘与湿地中的两栖动物物种进行监测（美国地质调查局2004年）。该方法面临的一大挑战是精准的物种识别，在无尾目动物（anuran）物种多样性极高的热带区域，这一问题尤为突出（如新热带区的无尾目动物物种丰富度高于新北区，Duellman 1999年）。此外，数量众多或鸣叫声洪亮的物种，可能会降低其他物种的被检测概率。声学调查中存在一项有待验证的假设：物种被检测到的概率与其存在与否或活动强度高度相关。若未对检测概率进行恰当校正，监测数据将无法准确反映物种的真实现状（MacKenzie等人2002年），进而可能导致不当的管理决策，造成时间与资金的浪费。数字记录仪（digital recorders）可作为长期监测项目中，两栖动物合唱群的物种检测与鸣叫活动强度评估的标准化手段。数字录音文件易于存储与传输，且可供专家在闲暇时开展录音分析。此外，可通过计算机软件（computer software）过滤鸣叫声洪亮或数量众多的物种，以提升其他物种的被检测概率。本研究验证了一项假设：相较于仅通过听觉分析录音，借助计算机软件分析录音时，无尾目动物合唱群的物种存在情况与鸣叫活动强度的检测准确率更高。检测准确率的提升，将有助于优化种群现状数据，并减少物种调查过程中的假阴性（false negatives）情况，即因干扰导致未检测到目标物种的问题。

搜集汇总

背景与挑战

背景概述

该数据集包含波多黎各三个地点（Carite Forest、Mount Britton、Tradewinds）的青蛙音频记录，用于研究数字录音机结合计算机软件分析对Eleutherodactylus青蛙物种检测效果的提升。数据集包括15个一分钟的wav格式音频文件和对应的ACI（声学复杂性指数）值，比较了仅听录音与软件分析后的检测结果，旨在改善两栖动物监测项目的准确性和减少误报。