Dataset: Middle Marsh 2014 Acoustic Time Series: raw files

Beginning in June of 2014, a short-term effort was initiated to record ambient underwater sound at high temporal resolution, within the Middle Marsh portion of the Rachel Carson Estuarine Research Reserve. This was accomplished using a low-power SoundTrap recorder and hydrophone (Ocean Instruments New Zealand).

The instrument was strapped vertically to a metal post and positioned ~0.15m above the seafloor, and in water of sufficient depth so that it remained submerged during spring low tides. At low tide, the SoundTraps were submerged at least 0.15 m, and at high tide, the SoundTraps were submerged in approximately 1 m of water, depending on the magnitude of the tide.

Three individual deployments were carried out using eight SoundTraps between 12 June 2014 and 31 August 2014. Deployment start and end dates for each site are listed in Table 1. The instrument was programmed to record for 130 seconds every 20 minutes at a 96 kHz sample rate. The SoundTrap analog signal is digitized at a fixed rate of 288 kHz. A digital anti-alias filter, with a cutoff frequency of 0.45 times the desired sample rate, is then applied before decimation. Consequently, at our sample rate of 96 kHz, the useable (-3 dB) bandwidth of these recordings is 0.020-43.0 kHz. In total, 36,870 separate 2-minute recordings were made as part of this monitoring initiative.

The data may be read into MATLAB using functions included in MATLAB.

Example:
% Set calibration
cal=10^(STcalibration/20);

[y,fs]=audioread(‘filename.wav’); %read in audiofile
y=y(fs*5:end) %remove first 5 seconds that have calibration sound
y=(y-mean(y))*cal; %demeaned and response correct to uPa

Additional information about this dataset can be found in Ricci et al (2016 and 2017).