Data from: "Proximate determinants of the frequency of mosquito sounds: separating species-specific effects from environmentally driven variations - implications for AI species recognition"

The dataset presented here contains the data and scripts of the paper "Proximate determinants of the frequency of mosquito sounds: separating species-specific effects from environmentally driven variations - implications for AI species recognition".Mosquitoes were collected from the fields as eggs or larvae, brought back to the laboratory, and bred until the imago stage. Then, we recorded free-flying individual mosquitoes in a small cage made of mosquito net, in a soundproof box, in the laboratory. Mosquitoes were recorded for 10 min, after which their sex and species were identified through morphological features, and the resulting sound was analyzed using Raven and R. Every individual was only tested once, but several sounds could be obtained for each recording. Here is the description of the different files:Mosquito_sound_data.txt:<br>- filename: The name of the audio file linked to the recording<br>- sound: Number of the sound in the recording. We aimed to obtain at least 10 good-quality sounds per recording, but in some cases, the number is lower<br>- frequency_mean: The mean frequency of the first harmonic (in Hz)- species: The species of the individual<br>- sex: The sex of the individual- individual: code of the individual mosquito<br>- origin: where the mosquito was collected (locality, if it was collected from the field, or colony origin if it came from an artificial colony)- temperature: temperature measured in the box during the recording (°C)<br>- rh: relative humidity measured in the box during the recording (%)- wing_standard: wing length as standardized per species<br>- wing_length: length of the individual’s wing (mm)- age_median: median age (in days) of the individual<br>- age_var: age variance in the breeding cage- age_min: minimum age (in days) of the individual (based on its breeding cage)<br>- age_max: maximum age (in days) of the individual (based on its breeding cage)- date_recording: date at which the recording took place (ordinal date)<br>- time_decimal: time of the recording- rec_system: recording system (custom 4-channel or Audiomoth)collection_data.txt: (data used for Fig1)<br>- Settlement: the city where the mosquitoes were collected<br>- species: how many species were collected in the settlement<br>- individuals: how many individuals were collected in the settlement<br>- lat: latitude of the collection point<br>- long: longitude of the collection point<br>Script mosquito sound analysis.R: <br>The script used for the analysis in the paper. <br>Script mosquito sound measurement.R:<br>The scripts used for the measurements of the mosquito sounds, and their visual validation using spectrograms. They require the .wav recordings and the .txt timestamps from the automatic mosquito sound detection (https://github.com/MosquiTUNE/Mosquito_sound_detection_methods/tree/main).<br><br>

本数据集包含论文《蚊子鸣声频率的近因决定因素：区分物种特异性效应与环境驱动变异——对AI物种识别的启示》的相关数据与分析脚本。
研究对象为野外采集的蚊卵或幼虫，将其带回实验室饲养至成虫阶段。随后于实验室隔音箱内的蚊帐小笼中，记录单只自由飞行蚊子的鸣声。每只蚊子的录音时长为10分钟，之后通过形态特征鉴定其性别与物种，并采用Raven和R软件对所得鸣声进行分析。每只个体仅接受一次测试，但单次录音可获取多段有效鸣声。

各文件说明如下：
1. Mosquito_sound_data.txt：
- filename：与该录音关联的音频文件名
- sound：该录音中的鸣声编号。本研究旨在单次录音中获取至少10段高质量鸣声，但部分案例中实际数量有所不足
- frequency_mean：第一谐波的平均频率（单位：Hz）
- species：该个体所属物种
- sex：该个体的性别
- individual：单只蚊子的个体编码
- origin：蚊子的采集来源：野外采集时为具体采样地点，人工饲养种群则为种群繁育来源
- temperature：录音期间隔音箱内的环境温度（单位：℃）
- rh：录音期间隔音箱内的相对湿度（单位：%）
- wing_standard：按物种标准化后的翅长
- wing_length：该个体的实际翅长（单位：mm）
- age_median：该个体的中位饲养天数
- age_var：对应饲养笼内种群的年龄方差
- age_min：该个体的最小饲养天数（基于其所在饲养笼的繁育记录）
- age_max：该个体的最大饲养天数（基于其所在饲养笼的繁育记录）
- date_recording：录音日期（以序数日期格式记录）
- time_decimal：录音的具体时刻（十进制格式）
- rec_system：录音设备系统（自定义4通道录音设备或Audiomoth）

2. collection_data.txt（用于生成图1的数据）：
- Settlement：蚊子的采集城市
- species：该采集点所获的蚊子物种数量
- individuals：该采集点所采集的蚊子个体总数
- lat：采样点的纬度
- long：采样点的经度

3. Script mosquito sound analysis.R：本论文数据分析环节所用的R脚本
4. Script mosquito sound measurement.R：用于蚊子鸣声测量及通过频谱图进行人工验证的R脚本。该脚本需依赖.wav格式的录音文件，以及来自自动蚊子鸣声检测工具的.txt格式时间戳文件（https://github.com/MosquiTUNE/Mosquito_sound_detection_methods/tree/main）