Tracking beaked whales with a passive acoustic profiler float
Implementation of ERMA on a DSP-based click detector. The digitized signal is stored in a 2-s buffer and band-passed sequentially in 10-ms increment at two frequencies f 1 and f 2. The Teger–Keiser operator is applied to the energy ratio e(n) between the two band-pass outputs to calculate the detection function d(n), which is indexed and compared with the detection threshold td . Only the clicks rising above the mean detection threshold are evaluated and used to estimate the percentage of clicks within the specified ICI range and ML.
(Color online) Block diagram of theQUEphone float system. Variable gain pre-amp's output is digitized by the DSP. The data are processed for click detections in real time and stored in a memory card. The float controller periodically checks the DSP for detection updates. The results are sent back to shore via Iridium® satellite and RUDICS server when it surfaces.
The AUTEC seafloor hydrophone array and the QUEphones' surface positions on each day. Q1 and Q3 were deployed near hydrophone H9 and H27, respectively, on June 7, 2010. They drifted at a depth of ∼1000 m towards H28 and were recovered near H28 on June 11, 2010 after moving approximately 6 and 3 km, respectively. Inset shows the estimated horizontal paths of the floats between known surface positions.
(Color online) Top curve (blue dash), Q3's ambient noise spectral density re 1 μPa2/Hz at AUTEC on June 9, 2010, 00:02 GMT (sea state ∼0) when it was drifting at 1005 m and bottom curve (red), SCORE's noise spectrum on January 5, 2011, 04:01 GMT (sea state 4) at ∼770 m when Q3 was still descending at ∼8 cm/s. 8-s long time series were analyzed by FFT with a 100-Hz moving average.
(Color online) (a) SPLRMS level of Q1 in the 25.75–28.25 kHz band in dB re 1 μPa2. (b) Beaked whale click counts for Q1 (yellow bars with red circles) and AUTEC fixed hydrophone H28 (blue bars). Dotted line (gray) is the depth of Q1 in meters. Horizontal line (brown) is the depth (450 m) at which the acoustic system was turned on/off. Dashed curve (pink) shows the approximate distance in meters between Q1 and H28.
(Color online) (a) SPLRMS noise level of Q3 in the 25.75–28.25 kHz band in dB re 1 μPa2. (b) Beaked whale click counts for Q3 (green bars with red circles) and AUTEC hydrophone H28 (blue bars). Dotted line (gray) is the depth of Q3 in meters. Horizontal line (brown) is the depth (450 m) at which the acoustic system was turned on/off. Dashed curve (pink) shows the approximate distance in meters between Q3 and H28.
(Color online) SNRs at 27 kHz with 2.5-kHz bandwidth of the click signals received by Q1 [red (crosses)] and Q3 [blue (stars)] during the AUTEC experiment.
(Color online) SNRs vs click counts of Q1 [red (crosses)] and Q3 [blue (stars)] in 248-s time bins. The highest SNR was 29 dB.
(Color online) (a) SNR vs percentage of ICIs within 0.2–0.5. (b) SNR vs mean likelihood (ML). Only the clicks whose ICI% are ≥30% and ML .
(Color online) Maximum ranges of beaked whale detection at receiver-source depth of 500 and 1000 m by ERMA detector as a function of sea states. Top blue (solid) and red (broken) lines are based on detection frequency at 27 kHz and bottom black (solid) and green lines (broken) are at 40 kHz.
Daily beaked whale's click counts by Q1, Q3, and the AUTEC M3R's count for hydrophone H28 from June 7 through June 11, 2010.
Article metrics loading...
Full text loading...