Soundscape dataset of Gregoire stream, located in southern Québec (2017)

Abstract: Data were collected in the summer 2017 in a second order stream with the objective to draw an underwater acoustic map at a high resolution for the entire stream (1150 meter). The sample scale was 2.5 meters. This dataset contains 917 sample sites that measured underwater acoustic profile in term of decibel (ref: 1uPa) per 1 hertz frequency (from 1Hz to 22000Hz). For each sample, the location in the stream is detailed. Acoustic profiles were made from underwater sound records of 10 s long. Method: The study was conducted in Gregoire Stream, a second-order stream in southern Quebec, Canada (46°05'40''N, 71°25'58''W). The area has a rain-snow regime characterized by peak water flows in spring and low flows in summer. We delineated a reach of 1150 m in length characterized by a flat low-flow section at the upstream end (stream distance = 1150 m) and a series of waterfalls at the downstream end (stream distance = 0 m), near to the confluence of Gregoire Stream and the Bagot River. The reach was crossed by a road which separated two zones differing in slope and riparian vegetation; a downstream zone having steeper slope and forest cover and an upstream zone having shallower slope and mostly shrub cover. Underwater soundscape was recorded along the entire stream reach (1150m). we recorded two 10-second underwater audio clip in each 2.5-m stream section (middle left and middle right). Underwater sounds were recorded using a H1 hydrophone powered by a PA4 amplifier (Aquarian Scientific, Anacortes, USA) and connected to a digital recorder (H2n Recorder, Zoom, Tokyo, Japan). A total of 917 10-second audio clips were recorded (16-bit WAV format at a sampling rate of 44.1 kHz) at a depth of 5 cm below the water surface. Each audio clip was listened with earphones to ensure that the hydrophone was not rubbing on hard surfaces while recording. The recording system has a maximum input of 10 V (rms) and a rated sensitivity of -190 dB re 1 V/µPa over the flat frequency-response range 10–3,000 Hz. Above specifications were used to obtain referenced sound pressure level (SPL) (dB re 1 µPa) with the PAMGuide functions (Merchant et al. 2015). For each audio clip, the frequency spectrum averaged through time was calculated to obtain referenced SPL values in 1-Hz frequency bands over the 0.1-10 KHz range. This procedure yielded a total of 22049 dB values per audio clip.