Aaron Thode

Matched-field processing, geoacoustic inversion, and source signature recovery of blue whale vocalizations

Matched-field processing (MFP) and global inversion techniques have been applied to vocalizations from four whales recorded on a 48-element tilted vertical array off the Channel Islands in 1996. Global inversions from selected whale calls using as few as eight elements extracted information about the surrounding ocean bottom composition, array shape, and the animals position. These inversion results were then used to conduct straightforward MFP on other calls. The sediment sound-speed inversion estimates are consistent with those derived from sediment samples collected in the area. In general, most animals swam from the east to west, but one animal remained within approximately 500 m of its original position over 45 min. All whales vocalized between 10 and 40 m depth. Three acoustic sequences are discussed in detail: the first illustrating a match between an acoustic track and visual sighting, the second tracking two whales to ranges out to 8 km, and the final sequence demonstrating high-resolution dive profiles from an animal that changed its course to avoid the research platform FLIP (floating instrument platform). This last whale displayed an unusual diversity of signals that include three strong frequency-modulated (FM) downsweeps which contain possible signs of an internal resonance. The arrival of this same whale coincided with a sudden change in oceanographic conditions.

Using ocean ambient noise for array self-localization and self-synchronization

Estimates of the travel times between the elements of a bottom hydrophone array can be extracted from the time-averaged ambient noise cross-correlation function (NCF). This is confirmed using 11-min-long data blocks of ambient noise recordings that were collected in May 1995 near the southern California coast at an average depth of 21 m in the 150-700 Hz frequency range. Coherent horizontal wavefronts emerging from the time derivative of the NCF are obtained across the arrays aperture and are related to the direct arrival time of the time-domain Greens function (TDGF). These coherent wavefronts are used for array element self-localization (AESL) and array element self-synchronization (AESS). The estimated array element locations are used to beamform on a towed source.

PAMGUARD: SEMIAUTOMATED, OPEN SOURCE SOFTWARE FOR REAL-TIME ACOUSTIC DETECTION AND LOCALISATION OF CETACEANS

PAMGUARD is open‐source, platform‐independent software to address the needs of developers and users of Passive Acoustic Monitoring (PAM) systems. For the PAM operator—marine mammal biologist, manager, or mitigator—PAMGUARD provides a flexible and easy‐to‐use suite of detection, localization, data management, and display modules. These provide a standard interface across different platforms with the flexibility to allow multiple detectors to be added, removed, and configured according to the species of interest and the hardware configuration on a particular project. For developers of PAM systems, an Application Programming Interface (API) has been developed which contains standard classes for the efficient handling of many types of data, interfaces to acquisition hardware and to databases, and a GUI framework for data display. PAMGUARD replicates and exceeds the capabilities of earlier real time monitoring programs such as the IFAW Logger Suite and Ishmael. Ongoing developments include improved real‐time l...

Probing Europa's interior with natural sound sources

Europas interior structure may be determined by relatively simple and robust seismo-acoustic echo sounding techniques. The strategy is to use ice cracking events or impacts that are hypothesized to occur regularly on Europas surface as sources of opportunity. A single passive geophone on Europas surface may then be used to estimate the thickness of its ice shell and the depth of its ocean by measuring the travel time of seismo-acoustic reflections from the corresponding internal strata. Quantitative analysis is presented with full-field seismo-acoustic modeling of the Europan environment. This includes models for Europan ambient noise and conditions on signal-to-noise ratio necessary for the proposed technique to be feasible. The possibility of determining Europas ice layer thickness by surface wave and modal analysis with a single geophone is also investigated.

Tracking sperm whale (Physeter macrocephalus) dive profiles using a towed passive acoustic array

A passive acoustic method is presented for tracking sperm whale dive profiles, using two or three hydrophones deployed as either a vertical or large-aperture towed array. The relative arrival times between the direct and surface-reflected acoustic paths are used to obtain the ranges and depths of animals with respect to the array, provided that the hydrophone depths are independently measured. Besides reducing the number of hydrophones required, exploiting surface reflections simplifies automation of the data processing. Experimental results are shown from 2002 and 2003 cruises in the Gulf of Mexico for two different towed array deployments. The 2002 deployment consisted of two short-aperture towed arrays separated by 170 m, while the 2003 deployment placed an autonomous acoustic recorder in tandem with a short-aperture towed array, and used ship noise to time-align the acoustic data. The resulting dive profiles were independently checked using single-hydrophone localizations, whenever multipath reflections from the ocean bottom could be exploited to effectively create a large-aperture vertical array. This technique may have applications for basic research and for real-time mitigation for seismic airgun surveys.

Localization using Bartlett matched-field processor sidelobes

Ambiguity surface sidelobes generated by the Bartlett matched-field processor (MFP) shift location with frequency. This sidelobe shift can be viewed as a continuous trajectory in a range-frequency plane at a fixed depth, where the trajectories converge to the correct source range for a perfectly matched surface. In isovelocity or bottom-interacting environments the sidelobe trajectories are straight lines that converge to the true range at zero frequency, while environments with upward-refracting sound-speed profiles have trajectories that asymptotically converge as the frequency approaches infinity. This behavior can be explained by the theory of waveguide invariants, which predict the local behavior of interference maxima/minima of acoustic intensity in the frequency-range plane. As the ambiguity surface of the Bartlett matched-field processor has a physical interpretation in terms of a time-reversed acoustic field, with the sidelobes analogous to local interference maxima, these invariant concepts can be reformulated for application to MFP. These interference trajectories are demonstrated to exist in simulations, broadband source tows, and a type A blue whale vocalization. Sidelobe trajectories also exist in the range-depth plane, but they contain no information about the correct source depth. An appendix demonstrates how these sidelobe properties can be exploited when combining ambiguity surfaces through use of gradient and Radon transform information. The resulting range estimators demonstrate better peak-to-sidelobe ratios than a simple incoherent average.

Quantifying seismic survey reverberation off the Alaskan North Slope

Shallow-water airgun survey activities off the North Slope of Alaska generate impulsive sounds that are the focus of much regulatory attention. Reverberation from repetitive airgun shots, however, can also increase background noise levels, which can decrease the detection range of nearby passive acoustic monitoring (PAM) systems. Typical acoustic metrics for impulsive signals provide no quantitative information about reverberation or its relative effect on the ambient acoustic environment. Here, two conservative metrics are defined for quantifying reverberation: a minimum level metric measures reverberation levels that exist between airgun pulse arrivals, while a reverberation metric estimates the relative magnitude of reverberation vs expected ambient levels in the hypothetical absence of airgun activity, using satellite-measured wind data. The metrics are applied to acoustic data measured by autonomous recorders in the Alaskan Beaufort Sea in 2008 and demonstrate how seismic surveys can increase the background noise over natural ambient levels by 30-45 dB within 1 km of the activity, by 10-25 dB within 15 km of the activity, and by a few dB at 128 km range. These results suggest that shallow-water reverberation would reduce the performance of nearby PAM systems when monitoring for marine mammals within a few kilometers of shallow-water seismic surveys.

Three-dimensional passive acoustic tracking of sperm whales (Physeter macrocephalus) in ray-refracting environments

A wide-aperture towed passive acoustic array is used to obtain ranges and depths of acoustically active sperm whales in the Gulf of Mexico in June 2004, by extending a technique previously reported [Thode, J. Acoust. Soc. Am. 116, 245-253 (2004)] to explicitly account for ray-refraction effects arising from a depth-dependent sound speed profile. Under this expanded approach, three quantities are measured from an impulsive sound: the time difference between direct-path arrivals on a forward and rear subarray, the time difference between the direct and surface-reflected paths on the rear subarray, and the acoustic bearing measured on the rear subarray. These quantities, combined with independent measurements of hydrophone depths and cable inclination, are converted into range-depth position fixes by implementing an efficient numerical procedure that uses a ray-tracing code to account for ray-refraction effects caused by depth-dependent sound speed profiles. Analytic expressions that assume a constant waterborne sound speed are also derived. Foraging depths of various sperm whales over 10 days in June, 2004 are estimated using the numerical technique.

Effects of Airgun Sounds on Bowhead Whale Calling Rates: Evidence for Two Behavioral Thresholds

In proximity to seismic operations, bowhead whales (Balaena mysticetus) decrease their calling rates. Here, we investigate the transition from normal calling behavior to decreased calling and identify two threshold levels of received sound from airgun pulses at which calling behavior changes. Data were collected in August–October 2007–2010, during the westward autumn migration in the Alaskan Beaufort Sea. Up to 40 directional acoustic recorders (DASARs) were deployed at five sites offshore of the Alaskan North Slope. Using triangulation, whale calls localized within 2 km of each DASAR were identified and tallied every 10 minutes each season, so that the detected call rate could be interpreted as the actual call production rate. Moreover, airgun pulses were identified on each DASAR, analyzed, and a cumulative sound exposure level was computed for each 10-min period each season (CSEL10-min). A Poisson regression model was used to examine the relationship between the received CSEL10-min from airguns and the number of detected bowhead calls. Calling rates increased as soon as airgun pulses were detectable, compared to calling rates in the absence of airgun pulses. After the initial increase, calling rates leveled off at a received CSEL10-min of ~94 dB re 1 μPa2-s (the lower threshold). In contrast, once CSEL10-min exceeded ~127 dB re 1 μPa2-s (the upper threshold), whale calling rates began decreasing, and when CSEL10-min values were above ~160 dB re 1 μPa2-s, the whales were virtually silent.

Automated detection and localization of bowhead whale sounds in the presence of seismic airgun surveys.

An automated procedure has been developed for detecting and localizing frequency-modulated bowhead whale sounds in the presence of seismic airgun surveys. The procedure was applied to four years of data, collected from over 30 directional autonomous recording packages deployed over a 280 km span of continental shelf in the Alaskan Beaufort Sea. The procedure has six sequential stages that begin by extracting 25-element feature vectors from spectrograms of potential call candidates. Two cascaded neural networks then classify some feature vectors as bowhead calls, and the procedure then matches calls between recorders to triangulate locations. To train the networks, manual analysts flagged 219 471 bowhead call examples from 2008 and 2009. Manual analyses were also used to identify 1.17 million transient signals that were not whale calls. The network output thresholds were adjusted to reject 20% of whale calls in the training data. Validation runs using 2007 and 2010 data found that the procedure missed 30%-40% of manually detected calls. Furthermore, 20%-40% of the sounds flagged as calls are not present in the manual analyses; however, these extra detections incorporate legitimate whale calls overlooked by human analysts. Both manual and automated methods produce similar spatial and temporal call distributions.