Acoustic recordings of bottlenose dolphins collected during boat-based encounters using HydroMoth devices and analysed with PAMGuard. This dataset contains 8 minutes of underwater acoustic recordings of a group of bottlenose dolphins (Tursiops truncatus), captured during a boat-based survey on 1 December 2024, approximately 5.3 nautical miles off the coast of Vilanova i la Geltrú, Spain, in the northwestern Mediterranean Sea. Recordings were made during cetacean encounters using a HydroMoth recorder (AudioMoth Firmware Basic v1.11.0), with the support of team members and volunteers from the Institut de Ciències del Mar (CSIC) and Associació Cetàcea. The survey lasted approximately 9 hours, from 2024-12-01T07:07:12 to 2024-12-01T16:51:00 (UTC). The first visual encounter with the dolphin group occurred at approximately 09:34:00 (UTC+1), after which the boat engine was turned off to increase the quality of the recordings. The audio files were analysed using the open-source software PAMGuard (version 2.02.16; DOI: 10.5281/zenodo.15056259). Whistles were classified using the ROCCA (Real-time Odontocete Call Classification Algorithm) delphinid whistle module integrated into PAMGuard. Clicks were manually categorised within the PAMGuard Viewer as echolocation clicks, burst-pulse sounds, or click trains. This dataset includes detailed metadata describing the entire event and the analysis, which encompasses up to 15 variables. The data are structured in accordance to Darwin Core standards: SamplingEvent, Occurrence, ExtendedMeasurementOrFact, and SimpleMultimedia. These key data are provided here and alongside the audio files in WAV format, available at https://doi.org/10.5281/zenodo.17175367.

[Datos del proyecto] Although the ROCCA model is capable of extracting up to 50 distinct acoustic features (see Oswald & Oswald, 2013, ROCCA User’s Manual), and the additional PAMGuard modules employed can generate over 300 supplementary acoustic parameters, only a targeted subset of variables relevant to this analysis has been retained in the final dataset published here. These selected variables focus on key click-related metrics (e.g., number of clicks, amplitude, and click type) and whistle-related characteristics (e.g., number of whistles, duration, mean frequency, and classification score). The variables were chosen based on their relevance to the study objectives and their suitability for downstream analyses. In total, the final subset comprises 15 variables, alongside standard metadata describing the sampling event and species occurrence, formatted according to the Darwin Core standard using the ExtendedMeasurementsAndFacts extension. These key data are provided alongside the audio files in WAV format, which are published at https://doi.org/10.5281/zenodo.17175367. Consequently, the acoustic recordings and their associated metadata are publicly available on Zenodo, while the complete set of over 300 measured variables, as well as all files containing the PAMGuard configuration—stored in the SQLite Databases and Binary Storage folders—are available upon request.

[Descripción de la metodología paso a paso:] All acoustic data recorded during the deployments were analysed both aurally and visually using the open-source software PAMGuard (version 2.02.16; DOI: 10.5281/zenodo.15056259) by biologists from the EMBIMOS research group at the Institut de Ciències del Mar (ICM-CSIC). The Whistle and Moan Detector and Click Detector modules were used to detect tonal and pulsed vocalisations, respectively. The ROCCA module (Real-time Odontocete Call Classification Algorithm) was employed to measure whistle contours, which were manually extracted using the contour-extraction graphical interface (see Oswald & Oswald, 2013, ROCCA User’s Manual). Spectrograms were visualised with the following settings: FFT length = 2048, hop size = 1024, and a Hann window with 50% overlap, resulting in a temporal resolution of approximately 8.19 ms. Additional modules and visualisation tools were configured to assist in the inspection of low-frequency acoustic events. For a complete description of the analysis workflow and parameters, refer to the PAMGuard Configuration Overview included with the dataset, which enables interpretation and replication of the study.

[Métodos de muestreo] The HydroMoth device was configured to capture high-frequency sounds at a sampling rate of 250 kHz, with medium gain. A band-pass filter ranging from 0.5 kHz to 125 kHz was applied. No amplitude threshold settings were enabled during the recording session. Upon sighting cetaceans and approaching them, the vessel’s engine was turned off to minimise acoustic disturbance, and the recorder was deployed at a depth of approximately 5 to 7 metres. The device remained submerged until the encounter was considered concluded—typically after several minutes without visual or acoustic detections, or when the animals were observed to be moving away or not returning. The recorder was then retrieved. Due to the high data volume generated by the recording settings (approximately 143 MB per 5 minutes at a 250 kHz sampling rate), the audio files were segmented into 1-minute clips prior to processing. This was done to reduce individual file sizes and enable their upload to open repositories—particularly those linked to citizen science initiatives—which often impose a maximum file size limit of 20 MB.

[Registros] Los datos en este recurso de evento de muestreo han sido publicados como Archivo Darwin Core(DwC-A), el cual es un formato estándar para compartir datos de biodiversidad como un conjunto de una o más tablas de datos. La tabla de datos del core contiene 1 registros. también existen 3 tablas de datos de extensiones. Un registro en una extensión provee información adicional sobre un registro en el core. El número de registros en cada tabla de datos de la extensión se ilustra a continuación. Event (core) 1 ExtendedMeasurementOrFact 120 Multimedia 8 Occurrence 8 Este IPT archiva los datos y, por lo tanto, sirve como repositorio de datos. Los datos y los metadatos del recurso están disponibles para su descarga en la sección descargas. La tabla versiones enumera otras versiones del recurso que se han puesto a disposición del público y permite seguir los cambios realizados en el recurso a lo largo del tiempo.

[Control de calidad] The organisations participating in this citizen science-based project were trained to use HydroMoth devices configured to optimise the detection of target species, based on local ecological knowledge and seasonal patterns. A standardised recording protocol was developed, and multiple training sessions were conducted for all participants involved in the data collection process. The resulting acoustic data were analysed by an experienced biologist from the EMBIMOS research group at the Institut de Ciències del Mar (CSIC). Quality control involved the careful configuration of PAMGuard modules to enable accurate automated detection, followed by manual review, annotation, and curation of the data using PAMGuard’s Viewer mode.

[Área de estudio] The study was conducted on 1 December 2024, during a boat-based survey in the northwestern Mediterranean Sea, approximately 5.3 nautical miles off the coast of Vilanova i la Geltrú, Spain. The survey lasted approximately 9 hours, from 07:07 to 16:51 UTC. Acoustic recordings were collected during a confirmed encounter with a group of bottlenose dolphins (Tursiops truncatus), observed for over 10 minutes. The recording effort focused on this specific encounter and its immediate surroundings, covering a localised area within the dolphins’ natural habitat.

This study was designed to collect underwater acoustic recordings of bottlenose dolphins (Tursiops truncatus) along the Catalan coast using HydroMoth devices during boat-based surveys. Data collection occurred during targeted encounters with dolphin groups, capturing vocalisations including whistles and clicks. Acoustic data were processed using the ROCCA model and PAMGuard software to extract relevant bioacoustic parameters. A subset of key variables was selected for analysis and publication, formatted following Darwin Core standards.

DTO-BioFlow – Integration of biodiversity monitoring data into the Digital Twin Ocean (Grant Agreement ID: 101112823).

Peer reviewed