Localization Of Whales Research Articles

Time synchronization and geoacoustic inversion using baleen whale sounds

In 1996 matched‐field processing (MFP) and geoacoustic inversion methods were used to invert for range, depth, and source levels of blue whale vocalizations. [A. M. Thode, G. L. D’Spain, and W. A. Kuperman, J. Acoust. Soc. Am. 107, 1286–1300 (2000)]. Humpback whales also produce broadband sequences of sounds that contain significant energy between 50 Hz to over 1 kHz. In Oct. 2003 and 2004 samples of humpback whale song were collected on vertical and titled arrays in 24‐m‐deep water in conjunction with the Humpback Acoustic Research Collaboration (HARC). The arrays consisted of autonomous recorders attached to a rope, and were time synchronized by extending standard geoacoustic inversion methods to invert for clock offset as well as whale location. The diffuse ambient noise background field was then used to correct for subsequent clock drift. Independent measurements of the local bathymetry and transmission loss were also obtained in the area. Preliminary results are presented for geoacoustic inversions of the ocean floor composition and humpback whale locations and source levels. [Work supported by ONR Ocean Acoustic Entry Level Faculty Award and Marine Mammals Program.]

Location and range of calling blue and fin whales off the Western Antarctic Peninsula

Blue and fin whales were heavily whaled from the Southern Ocean during the 20th century and their populations today remain at extremely low levels. Both species produce loud, low-frequency calls that are well suited for long propagation. Eight acoustic recording packages (ARPs) were deployed off the Western Antarctic Peninsula, with instrument spacing from 100 to 180 km. Blue and fin whale calls were commonly heard and two different methods were used to determine locations and ranges to the calling whales. Blue whales were localized using the differences in call arrival times to three different instruments. It also was possible to calculate ranges to blue whales from multipath arrivals of their calls and verify the results obtained by the difference in arrival times method. Short detection ranges for fin whales precluded localization, but multipath arrival difference allowed range estimation. Source levels for both species were calculated using these ranges and the measured received levels.

Acoustic localization of bowhead whales near a Beaufort Sea oil development, 2001–2003: Evidence of deflection at high-noise times?

Most bowhead whales migrating west along Alaska’s north coast during autumn pass 10–75 km seaward of BPs Northstar oil production island. Our objective was to determine the extent to which underwater sound from Northstar deflected the closest whales farther offshore. Whale calls were localized, via 11 DIFAR-based seafloor recorders, on 23–36 days per year while underwater sound was measured continuously 450 m from Northstar. Totals of 1121, 2057 and 14<th>304 calls were localized within a 20×26 km analysis area seaward of Northstar. Quantile regression was used to relate the 5th quantile distance-from-shore to underwater sound. A permutation method allowed for lack of independence of closely spaced calls. In 2001, the closest calling bowheads tended to be slightly farther offshore at times with high industrial sound, which were mainly due to vessels near Northstar. The displacement effect was maximal if sound was averaged over 50–80 minutes preceding calls. In 2002, high sound levels were rare and results were inconclusive. In 2003, offshore displacement was not significant despite high sample size and a migration corridor unusually close to shore. The upper 90% confidence bound on the percentage of bowheads displaced 2+ km did not exceed 1.6% in any year. [Work supported by BP.]

The use of acoustic multipath for localization of sperm whales

Sperm whales produce impulsive sounds, or ‘‘clicks,’’ that are ideally suited for acoustic tracking. If a widely distributed passive acoustic array is available, then the relative time delays from a click’s direct path can be used to derive an animal’s location. There are many interesting circumstances, however, where deploying a wide-baseline 3D array geometry is either unavailable or impractical. In such situations acoustic reflections of clicks from the ocean surface and bottom can be used to effectively convert a few hydrophones into a wide-aperture array that permits localization in range, depth, and azimuth. In addition acoustic multipath, particularly surface reflections, can also simplify automated detection and tracking procedures, and permit a single hydrophone to be used to distinguish between multiple whales calling simultaneously. Reviewed here is recent research on how multipath effects detected on one or two hydrophones have been used to recover dive profiles of sperm whales in the Gulf of Alaska and the Gulf of Mexico. This information, in turn, has been used to explore whether the animals exploit multipath for their own purposes, to estimate signal directionality, and to observe how sperm whales remove fish from longlines. [Work sponsored by ONR, the Minerals Management Service, and the North Pacific Research Board.]

Directional frequency and recording (DIFAR) sensors in seafloor recorders to locate calling bowhead whales during their fall migration.

Bowhead whales, Balaena mysticetus, migrate west during fall approximately 10-75 km off the north coast of Alaska, passing the petroleum developments around Prudhoe Bay. Oil production operations on an artificial island 5 km offshore create sounds heard by some whales. As part of an effort to assess whether migrating whales deflect farther offshore at times with high industrial noise, an acoustical approach was selected for localizing calling whales. The technique incorporated DIFAR (directional frequency and recording) sonobuoy techniques. An array of 11 DASARs (directional autonomous seafloor acoustic recorders) was built and installed with unit-to-unit separation of 5 km. When two or more DASARs detected the same call, the whale location was determined from the bearing intersections. This article describes the acoustic methods used to determine the locations of the calling bowhead whales and shows the types and precision of the data acquired. Calibration transmissions at GPS-measured times and locations provided measures of the individual DASAR clock drift and directional orientation. The standard error of the bearing measurements at distances of 3-4 km was approximately 1.35 degrees after corrections for gain imbalance in the two directional sensors. During 23 days in 2002, 10,587 bowhead calls were detected and 8383 were localized.

Passive tracking of multiple diving sperm whales using single hydrophones at two mobile locations

A simple method using only two hydrophones, each loosely deployed from separate mobile platforms, has been developed to simultaneously track several vocalizing sperm whales. The separation distance of several hundred meters between the hydrophones implies that a particular whale could present vastly different beam orientations towards them, thus precluding use of any specific relationship between the click signal levels at the two locations for the purpose of whale localization. This method utilizes time-of-arrival of the direct clicks and their surface reflections and matches them at the two hydrophone locations. Whales are segregated on the basis of the set of observed time delays. Click parameters such as interpulse interval and average click frequency are proposed to be used as secondary data for the purpose of track refinement. Depth profile of the whales is obtained independent of the separation distance between the two hydrophones. However, knowledge of the separation distance between the hydrophones provides the 3D coordinates of the whales within a left–right ambiguity. Whale tracks have been obtained using actual data collected off Ogasawara Islands in Japan. These results demonstrate the utility of this method for studying the bioacoustics and behavior of deep diving sperm whales.

Historical Population Characteristics of Bowhead Whales (<i>Balaena mysticetus</i>) in Hudson Bay

Historical records of commercial whalers operating in northwestern Hudson Bay during the 19th century were examined for information on size, age, sex, and location of bowhead whales that were either sighted or killed. Correlations between body size and either oil yield or baleen length were used to estimate the relative age classes (calf, subadult, adult) of whales for which no explicit age-class information was reported in the whaling logbooks. Cow-calf pairs and subadults, as well as adult whales, were sighted or killed throughout the whaling season in the area extending from Wager Bay south to Marble Island. This finding indicates that whales of many different age classes were present south of Wager Bay, even during the open-water period when whaling activity shifted northward to include Repulse Bay and Lyon Inlet. Recent observations suggest that few bowhead whales occur south of Wager Bay during the open-water season and that the population in this area has not recovered from the effects of commercial whaling. It is not clear whether this group of bowheads was a separate stock or, alternatively, waters south from Wager Bay constituted a second calf-rearing area for a single Hudson Bay-Foxe Basin stock.

Model-based tracking of marine mammals near California using seismometers

An algorithm originally developed for tracking humpback whales around a deep-water hydrophone array near Hawaii has been proven capable of localizing another species of marine mammal in the shallow waters off California. The new sources of interest are blue whales, animals with markedly different call characteristics from humpback whales, and the data under examination is from four bottom-mounted seismometers in a 3 km square array. The algorithm uses a range-dependent acoustic model to predict time differences of arrival (time-lag) of blue whale calls as measured between sensor pairs, while real pairwise time-lags are measured through a phase-only correlation process. Comparison between modeled and measured time-lags forms an ambiguity surface identifying the most probable whale location in a horizontal plane around the sparse array. The robustness of the model-based localization technique is illustrated in its application to a different scenario than for which it was developed, and it is also suitable for continuous, real-time, unattended alert and tracking applications. Examples of tracking whales along their migratory path will be provided.

ATOC AND OTHER FACTORS AFFECTING THE DISTRIBUTION AND ABUNDANCE OF HUMPBACK WHALES (MEGAPTERA NOVAEANGLIAE) OFF THE NORTH SHORE OF KAUAI

AbstractHumpback whale diving behavior changes subtly when exposed to signals transmitted from the Acoustic Thermometry of Ocean Climate (ATOC) sound projector located 14 km offshore the island of Kauai. This study considered whether such responses would lead to changes in distribution and abundance. A land‐based shore station measured humpback whale locations (scan samples) for both nearshore (<5 km) and offshore (5–10 km) areas. Control observations were made in 1994 and 1998. In 1998 multipleday blocks with ATOC transmissions were interspersed with multiple‐day control blocks without transmissions. Sighting rates were higher in 1998 (ATOC) than in 1994 (control year), probably due to better sighting conditions, but may reflect increased population size. Sighting rates did not differ between control and ATOC conditions in 1998. A seasonal sighting peak was observed in both years. No vessel effect on sighting rate was detected in 1998.There was no effect of ATOC on the distance from the shore station to whales, or the depth of water where pods were located. However, the distribution of whales shifted slightly eastward during the ATOC blocks and the mean distance between the ATOC source and pods was greater during transmissions. Nonetheless, more whales were found close to the source when it was on, suggesting a more variable response rather than simple avoidance, with whales found both closer to, and farther away from, the source during transmissions.

ESTIMATED SOURCE LEVELS OF FIN WHALE (BALAENOPTERA PHYSALUS) VOCALIZATIONS: ADJUSTMENTS FOR SURFACE INTERFERENCE

AbstractWe recorded a series of 34 fin whale vocalizations using four calibrated hydrophones in a towed array. The whale's location was determined using arrival‐time differences of each call at the hydrophones. The recordings showed evidence of interference between signals propagating along direct and surface‐reflected paths (the “Lloyd Mirror”). We made preliminary estimates of source level (SL) using a spherical spreading model of transmission loss. We then applied a conservative correction for the effect of interference by subtracting the maximum possible signal enhancement from the preliminary SL estimate. These adjusted spherical spreading estimates had a median of 171 dB re 1 μPa at 1 m, with a range of 159–184 dB. These are the first SL estimates that explicitly attempt to correct for possible interference effects. We also propose a method that simultaneously estimates SL and source depth by fitting a Lloyd Mirror transmission loss curve to the measured received level, known receiver depth, and source‐to‐receiver range. Attempts to apply this method to the present data set yielded inconsistent results, probably because of the small number of hydrophones available. With more hydrophones in the interference field, this new method should produce reliable estimates of SL and depth of vocalizing whales.

Automated model-based localization of marine mammals near Hawaii

Acoustic data from the Pacific Missile Range Facility (PMRF) hydrophone network near Hawaii is being used to develop a real-time automated alert and tracking algorithm. The sources of interest are marine mammals, humpback whales in particular, which are regularly seen in the vicinity of the PMRF array. The algorithm under development uses acoustic data from six hydrophones plus an acoustic propagation model to construct an ambiguity surface identifying the most probable whale location in a horial plane around the array. It has the further advantage that it can be implemented in real-time and without human interaction, making it suitable for automated alert applications.

Three-dimensional localization of diving sperm whales using a short-aperture towed horizontal array

Three-dimensional sperm whale localizations are typically obtained via recording their ‘‘clicks’’ on widely distributed hydrophones. Here, an alternative 3D localization method is investigated, using data collected by an 8-m aperture, five-element horizontal towed array deployed from the NOAA ship Gordon Gunter in the Gulf of Mexico, as part of a recent sperm whale pilot study. During the night of 3 July 2000, the Gunter towed the array at a steady 1-kn speed and 60-m depth through a pod of sperm whales, in 1000-m-deep water. During that time, surface and bottom reflections from a single sperm whale click were often recorded. By measuring the bearings and relative arrival times of the direct arrival and reflections, the whale location, array depth, and array tilt could be computed. The latter results were checked against measurements obtained from a time-depth recorder attached to the array. Assuming that the speed of the ship remained fixed and the prevailing currents did not change, subsequent positions could then be obtained using only a surface reflection. This method may provide a convenient way for observing dive profiles of animals within a few kilometers of a survey ship. [Work supported by Minerals Management Service, National Marine Fisheries Service, and ONR.]

Three-dimensional localization of blue whales using matched-field processing techniques: Unusual behavior, unanswered questions

Recently, matched-field processing (MFP) and global inversion methods were used to track the three-dimensional trajectories of six blue whales, during a 1996 experiment off San Miguel Island in California, conducted off the research platform FLIP, using a 48-element, tilted vertical hydrophone array. Once the animals’ positions and the propagation environment were determined, source levels were obtained for individual animals, and potential propagation effects were removed from their sounds. Out of the six whales studied, one showed unusual behavior. Initially it swam directly towards FLIP, while producing a diverse set of calls that included three strong FM sweeps. It then fell silent, altered its course to avoid FLIP, and began calling again as it swam away. Its closest point of approach to FLIP coincided with the arrival of a powerful tidal bore. The maximum source levels estimated from this animal were 10 dB lower than those from the other animals, and the FM sweeps showed signs of a possible resonance, which remained even when possible propagation effects were removed. The ability to precisely position blue whales has raised interesting questions about baleen whale sound production, the purpose behind these sounds, and the potential association between feeding animals and internal waves. [Work sponsored by ONR.]

Objective Functions for Ocean Acoustic Inversion Derived by Likelihood Methods

Selection of a suitable objective function is an integral part of the inverse problem, and poor selection can have a strong influence on the inverse result. Objective functions are here derived for many practical occasions such as for single frequency and broadband, with and without knowledge of source strength, and with and without the received signal phase. These objective functions are all derived from a unified approach based on maximum likelihood and additive Gaussian noise models. The assumptions for the objective function are thus clear and the resulting estimator has good properties. From a Bayesian point of view, the solution to the inverse problem is the a posteriori probability distribution of the unknown parameters, which can be found from the likelihood function. Thus using objective functions based on likelihood functions facilitates computing the a posteriori distributions.

Detection and censusing of blue whale vocalizations along the central California coast using a former SOSUS array

To investigate the feasibility of automating the detection and censusing of blue whale vocalizations over a large coastal region using the Naval Postgraduate School (NPS) Ocean Acoustic Observatory (OAO), a four-day experiment was conducted along Central California in the summer of 1997. During the experiment, array data were archived continuously at the NPS OAO, a former SOSUS array. In addition to shore-based acoustic monitoring, an aircraft was assigned to locate blue whales in the Monterey Bay National Marine Sanctuary, and a research vessel, manned with observers and instrumented with a towed hydrophone array, was used to confirm locations of the blue whales and classify the vocalized near-field signals. The shipboard measurements were required to provide a means to separate the source signal characteristics from their multipath signatures. The towed array data were deverberated, source level and characteristics were estimated, and call-to-call variability was studied. Based on robust signal parameters, two simple autodetection correlation filters for the ‘‘A’’ and ‘‘B’’ blue whale calls were constructed, respectively. By applying the autodetection filters to the OAO data, the number of blue whale calls for the summer of 1998 was tallied. [Research supported by SERDP/ONR.]

Techniques for passive detection and localization of marine mammals using wide aperture arrays

SACLANTCEN is sponsoring a program entitled Sound Oceanography and Living Marine Resources Program, whose objective is to demonstrate dual use technology for passive detection and localization of marine mammals, which will assist in the implementation of acoustic risk mitigation policies when using high-power acoustic sources. A sea trial was conducted in August 1999 to evaluate marine mammal detection and localization techniques using a number of advanced passive arrays. Passive, bearing-only data from marine mammal vocalizations received on a towed horizontal line array and target motion analysis are used to localize whale position. The multipath arrival structure of received signals and a ray trace model are used to determine whale range and depth over time. Data from several stations, executed using three deployed vertical line arrays in the vicinity of diving whales, and triangulation of arrivals are used to track whale location in the horizontal plane, while beamforming of the vertical array data is used to track whale depth. Evaluation of these data and techniques will be used to formulate passive procedures for determining the presence of marine mammals prior to activation of high-power acoustic sources.

Acoustic census of sperm whales in the eastern temperate North Pacific

An NOAA/ONR acoustic line-transect survey was conducted in spring 1997 to estimate the abundance of sperm whales in the eastern temperate North Pacific using a 52-m research vessel. An ITI hydrophone array was towed at 100-m depth as the vessel surveyed approximately 25 000 km of predetermined transect lines at 8 kn. Acoustic signals received by a pair of hydrophones were digitized and sent up to the ship via a coaxial tow cable. Bearing angles to the locations of sperm whales were calculated in real time based on differences in signal arrival times between the two hydrophones, and whale locations were determined by the convergence of bearing lines as the ship continued along its course. The ship was maneuvered to resolve the left/right ambiguity in whale locations and, during daylight hours, was directed to this estimated location to obtain visual estimates of group size when the whales surfaced. A total of 102 distinct groups of sperm whales were acoustically detected with sufficient amplitude to estimate bearing lines. ‘‘Slow clicks’’ (with a period of 4–8 s) were detected at ranges up to 20 nmi, whereas the more typical sperm whale clicks (with a period of 0.5–1.0 s) were typically heard at less than 5 nmi.

A feasibility field study of monitoring blue whales using the Pt. Sur Ocean Acoustic Observatory

In the summer of 1997, two three-day experiments were conducted to test the feasibility of acoustically detecting, classifying, localizing, and tracking blue whales at long ranges using a former SOSUS listening array located at the Naval Postgraduate School Ocean Acoustic Observatory (OAO) at Pt. Sur, California. During each experiment, full-array data were archived continuously at the OAO. In concert with the shore-based acoustic monitoring, an aircraft was assigned to locate blue whales in the Monterey Bay National Marine Sanctuary and to direct a research vessel to a whale site. The research vessel was manned with observers and instrumented with a towed hydrophone array to ground-truth the locations of the blue whales and classify the vocalized near-field signals. These shipboard measurements were required to provide a means to separate the source signal characteristics from the multipath signatures for the calibration and validation of broadband, model-based localization methods. In this presentation, an overview of this feasibility field study is provided. Initial experimental as well as modeling results are presented, which include assessments of the predictability (i.e., variability) of the vocalized sound and the uniqueness of the location-dependent multipath structure, both are fundamental to the applicability of model-based algorithms. [Research supported by ONR.]

The neutral network analysis of false killer whale (Pseudorca crassidens) vocalizations

This study reports the use of unsupervised, self-organizing neutral networks to categorize the repertoire of false killer whale vocalizations. Self-organizing networks are capable of detecting patterns in their input and partitioning those patterns into categories without requiring that the number or types of categories be predefined. The inputs for the neural networks were two-dimensional characterizations of false killer whale localizations where each vocalization was characterized by modulations in duty cycle and peak frequency [Murray et al. (in prep.)]. The first neural network used competitive learning, where neurons in a competitive layer distributed themselves to recognize frequently presented input vectors. This network resulted in classes representing typical patterns in the vocalizations. The second network was a Kohonen feature map which organized the outputs topologically, providing a graphical organization of pattern relationships. The networks performed well as measured by (1) the average correlation between the input vectors and the weight vectors for each category, and (2) the ability of the networks to classify novel localizations. The techniques used in this study could easily be applied to other species and facilitate the development of objective, comprehensive repertoire models.

Satellite-Monitored Movements of the Northern Right Whale

The northern right whale, Eubalaena glacialis, remains the most critically endangered of the large cetaceans despite international protection since 1936. We used satellite-monitored radiotags to identify the late-summer and fall habitat use patterns of right whales in the western North Atlantic. We tagged 9 whales in the Bay of Fundy (BOF) and successfully tracked them for a total of 13,910 km (x = 1,546 km) in 195 whale-tracking days (range 7-42 days each, x = 21.7 days). Individuals tracked for more than 12 consecutive days (N = 6 whales) left the BOF at least once and had higher average speeds (x = 3.5 km/hr) than those that stayed within the bay (x = 1.1 km/hr). Three of the tagged whales not only left the BOF, but traveled more than 2,000 km each before returning to the general tagging area. One adult female with a calf went to New Jersey and back to the BOF (3,761 km) in 42 days. Most locations were along bank edges, in basins or along the continental shelf. Eighty percent of locations were in water <182 m (100 fathoms [F]) deep. All of the tagged whales were located in or near shipping lanes. Right whale distribution coincided with areas intensively used by humans for fishing, shipping, and recreation. Individuals moved rapidly among areas previously identified as right whale habitat. Whale locations plotted on sea surface temperature (satellite infrared) images suggest that one whale spent time at the edge of a warm core ring and others spent extended periods in upwellings. Observations of whales surfacing with mud on their heads suggest that these whales fed near the BOF seafloor. Satellite telemetry is a useful means of tracking cetacean species that are difficult to view, move long distances, and might be too expensive to monitor by other means.