Locations of 12 autonomous hydrophone moorings (stars) moored along the Mid-Atlantic Ridge, and approximate locations (dotted boxes with circles) of seismic airgun activity located via the array.
(Color online) (a) Spectrogram and time of series of airgun pulses recorded 19 August 2002 on the 26° N 50° W hydrophone (spectrogram parameters: frame and FFT length 4.7 s (512 samples) overlap 0.75, Hamming window, for a filter bandwidth of 0.9 Hz). (b) Spectrogram and time series of fin whale 20 Hz pulses recorded 05 November 2005 on the 32° N 35° W hydrophone (spectrogram parameters: frame and FFT length 8.2 s (2048 samples) overlap 0.75, Hamming window for a filter bandwidth of 0.50 Hz). (c) Example of co-occurrence of airgun and fin whale sounds. In instances where the fin whale sounds were of lower amplitude than those in this figure, fin whale pulses would be obscured by airgun noise. (d) Example of airgun pulses (spectrogram parameters: frame and FFT length 2.3 s (256 samples) overlap 0.75, Hamming window, for a filter bandwidth of 1.8 Hz) recorded over 3900 km from the source. These sounds were recorded on the 32° N 35° W hydrophone; the survey vessel producing them was acoustically located in Brazilian waters.
(Color online) Example calculation of the fin index: (a) Log-scaled spectrogram S(t,f) of the sound (spectrogram frame and FFT size = 1 s, overlap = 0.5 s, Hann window, for a filter bandwidth of 4.0 Hz) and then averaged over 1-day periods; (b) normalized spectrogram Š(t,f) (at each time step t, the noise floor n 50(t) was subtracted from the spectrogram at that time step, with negative values converted to 0); (c) resulting fin index, or relative estimate of fin whale calling; (d) levels of sound in the adjacent frequency bands (upper: 40–45 Hz; lower: 8–13 Hz), used to ensure transient noise was not affecting the fin whale index.
Seasonal patterns of airgun pulses detected in data at the 12 mooring sites. Black bars represent percentage of days/month that airgun pulses were detected. Gray bars indicate periods for which there were no data available.
Seasonal patterns of fin whale 20 Hz pulses detected in data at the 12 mooring sites. Black dots represent the calculated fin index, or relative estimate of fin whale calling. Gray bars indicate periods for which there were no data available.
Normalized annual fin index peak. For each site, data were normalized by dividing the annual peak index by the maximum peak index over all years. Linear regression trend line: R2 = 0.47, p < 0.001. The positive trend indicates that fin whale calling increased over the duration of this study.
(Color online) Seasonal patterns of airguns (black bars, left y axis) and fin index (light gray line, right y axis) for sites in the southern array. Dark gray bars indicate periods for which there were no data available.
Article metrics loading...
Full text loading...