Passive hydroacoustic dataset collected at the shallow hydrothermal field of Vulcano Island (Southern Italy)

The repository contains a passive acoustic dataset acquired through multiple short‑term deployments (up to ~10 h each) performed at the shallow hydrothermal field of Baia di Levante (Vulcano Island) between 2022 and 2024. Acoustic data were collected using icListen SA9L‑ETH Smart Hydrophones, acquiring high‑resolution audio frames at 32 kHz. During each deployment, a hydrophone was moored close to a mid/high‑flow submarine gas seep at a distance of ~1 m. The repository includes all specific information for each deployment, such as date and time of acquisition, geographic position, depth, instrument model and serial number, sampling rate, and notes on operational conditions. The repository provides Sound Pressure Levels (SPL) in 1/3‑octave bands, broadband SPL in two frequency ranges (80–2500 Hz and 10–12800 Hz), which correspond to bubbling activity spectral signals and the hydrophone bandwidht, respectively. In the latter, frequency range 1-10 Hz was not considered to avoid potential low-frequency noise.For each audio frame, SPL was obtained by converting Power Spectral Density (PSD) values using hydrophone specifics (e.g., sensitivity, gain, peak‑to‑peak voltage); PSDs were computed using Welch’s method with 5‑second Hanning windows and a 50% overlap. It is worth noting that only SPL extracted from quality‑validated acoustic samples considered. They were selected after a dedicated filtering and inspection workflow to ensure that each record represents reliable hydrothermal activity and ambient noise conditions. Furthermore, deployment metadata and gas‑composition measurements are supplied in MS Excel format, while representative raw audio excerpts from each deployment are provided as WAV files. For each deployment, we provide figures showing the averaged PSD, the 1/3-octave SPL distribution, and a representative spectrogram derived from a 120‑s audio frame. All figures are supplied in PNG format and included in the Figures.zip file. This dataset provides a reusable resource for studies of hydrothermal dynamics, inverse modelling, ambient noise characterisation, and machine‑learning applications in underwater acoustics.