Science on Whales, Submarines, and Conservation Meets the News


In the jargon of acoustic signal processing, we each see the world through our own filter. Recent research into the effect of sonar on marine mammals provides a case study of how research results become erroneous perceptions that are reinforced and expanded in the popular press. Biases impede an informed discussion about research findings, and results are filtered and pulled out of context. It should not be a surprise that results of research on the effects of sonar on marine mammals, animals that hold an almost mythical place in the public arena, would suffer such a fate.

It is particularly important to recognize the effect of both research and media bias when study results are translated directly into management policy. Restrictions on the use of sonar during navy training exercises ultimately may mean the difference between life and death for men and women in uniform. For the U.S. Navy, with a defense strategy that revolves around the carrier-battle group (i.e., an aircraft carrier surrounded by an array of support ships), the threat of quiet diesel–electric submarines is not an abstraction.

The diesel–electric submarine's history of lethality stretches from World War I to the present. The case of the torpedoed Korean warship Cheonan, which in 2010 carried 46 sailors to their graves, is a grim reminder. Iranian threats to block commercial traffic from the Straits of Hormuz represent the latest danger and make ecologically responsible training of antisubmarine crews a priority for western navies.

Diesel–electric submarines, once submerged, cannot be detected by listening for them with passive sonar. Surface-ship defenses must instead rely primarily on midfrequency active (MFA) sonar, in which sound below 10 kHz is transmitted in an attempt to detect its reflection off the stealthy adversary. The need to train with sonar is juxtaposed against mass strandings of marine mammals coincident with MFA sonar operations (Frantzis 2003; D’Amico et al. 2009). A mass stranding is defined as at least two animals moving onto a beach, excluding a mother–calf pair. Perhaps the most studied stranding event associated with MFA sonar testing occurred on 15 March 2000 in the Northwest Providence Channel in the Bahamas. During a multiship sonar operation, 3 Blainville's beaked whales (Mesoplodon densirostris), 11 Cuvier's beaked whales (Ziphius cavirostris), and 2 minke whales (Balaenoptera acutorostrata) were stranded over a distance approximately 100 km long (Balcomb & Claridge 2001; England et al. 2001). In response to such events, research over the past 10 years has focused on determining the mechanisms by which anthropogenic noise, including sonar, may affect the behavior of marine mammals, especially beaked whales.

Cuvier's and Blainville's beaked whales are the Olympian divers of the cetacean world. Individuals of each species routinely dive every 1–2 hours to depths in excess of 1000 m. The animals often remain submerged for nearly an hour and spend little time at the surface. They produce echolocation clicks only at depth during these dives and thus the detection of these clicks can be used as a means of measuring their distribution and behavior in time and space (Zimmer et al. 2005; Baird et al. 2006).

The U.S. Navy operates large undersea ranges for tracking of subsurface vehicles, including submarines, torpedoes, and targets. The ranges are instrumented with many widely spaced, bottom-mounted hydrophones (underwater microphones). The hydrophones are being used to develop algorithms for passive acoustic detection, classification, localization, and density estimation of marine mammals (Moretti et al. 2002). This research is being conducted at three major facilities including the Atlantic Undersea Test and Evaluation Center (AUTEC) in the Tongue of the Ocean (TOTO) in the Bahamas. The TOTO forms the southern branch of the Grand Bahama Canyon and joins the Northwest Providence Channel, the site of the 2000 stranding of beaked whales. Passive acoustic monitoring of AUTEC hydrophones and visual observations have established that beaked whales are consistently present in the area (Marques et al. 2009; Moretti et al. 2010). More recent data at the Southern California Offshore Range show the presence of Cuvier's beaked whales (Falcone et al. 2009). Multiship MFA sonar operations are conducted at both ranges.

Data collected at AUTEC strongly suggests beaked whales are sensitive to anthropogenic sound. In April 2010, two articles describing the reaction of Blainville's beaked whales were published (McCarthy et al. 2010; Tyack et al. 2010). Both documented the reaction of Blainville's beaked whales at AUTEC to sound, including sonar. McCarthy et al. (2010) found Blainville's beaked whales move off range in response to sonar and return days after the cessation of the operations. In 2007, Tyack et al. (2010) exposed a beaked whale on two successive deep-foraging dives to a sonar-like signal and killer whale (Orcinus orca) calls. A Blainville's beaked whale was exposed in 2008 to a pseudo-random noise with the same duration and within the same frequency band as both the sonar and killer whale signals used in the 2007 experiment. In each case, the animal stopped foraging and ascended to the surface at a significantly slower rate than similar animals that were not exposed to such signals. A Blainville's beaked whale, to which a satellite tag was attached to in 2009, moved from the northern edge of the range ahead of a multiship sonar operation, moved farther north of the range during the operation, and returned to the range after the operation was complete. The results of these studies strongly suggest Blainville's beaked whales react to anthropogenic sound at a relatively low sound level.

Media Response

The media response to the papers by McCarthy et al. (2010) and Tyack et al. (2010) began on 14 March 2011 with a measured story in the BBC online publication Earth News, “Beaked Whales ‘scared’ by Navy Sonar” (Davies 2011). The evolution of this story presents an interesting and cautionary example of the portrayal of scientific reports in the popular press. In a series of articles following the BBC publication, sonar not only “scared” whales but also “chased whales away” (Mulvaney 2011) and “drove terrified beaked whales to their death” (Derbyshire 2011; Mulvaney 2011). According to the The Telegraph, the research on underwater sonar “proved” wind farms strand whales (Johnson 2011). Although the latter story was retracted on 17 March 2011 in response to a scathing letter from one of the principal investigators, the retraction mattered little to Newsweek, which subsequently reported, “A new study suggests that beaked whales—the size of a rhino and the weight of a bus—are confused by the presence of wind farms at sea, leading them to beach” (Newsweek 2011).

The rhinoceros reference, which first appeared in The Telegraph article on 15 March, provides insight into how concepts are iterated from article to article. The reference appeared in a Mail Online article later that day and again in Newsweek on 20 March.

Quotations from principal investigators were liberally transplanted from article to article. For example, the following first appeared on 14 March in the original Earth News article (Davies 2011): “It was clear that these whales moved quickly out of the way of the [navy] sonars. We now think that, in some unusual circumstances, they are just unable to get out of the way and this ends up with the animals stranding and dying.” The quotation was repeated with slight omission in a Discovery News article on 23 March (Cannon 2011).

A myriad of valid questions are raised by the original research, none of which were discussed by the media. For example, if animals react to sonar, why are they routinely detected on U.S. navy ranges, areas of repeated multiship MFA sonar operations? If a multiship sonar operation stranded animals in the Northwest Providence channel in 2000, why have no mass strandings been observed adjacent to the AUTEC range? Does habituation affect the severity of the whales’ response and, if so, do animals in the Bahamas move back and forth between the TOTO and the Northwest Providence Channel? Were stranded animals in the Northwest Providence Channel in 2000 naïve to sonar or exposed to sonar in the past? What is the effect on beaked whale populations of displacement during sonar operations?

Meeting the challenge of balancing sound management policy with the prescient need to train personnel in the use of active sonar requires a clear understanding of the mechanisms that drive whale strandings. The press will continue to report events that are biased by a predetermined story line and titled with an editor's catch phrase. Ionnidis (2005) and Freedman (2010) point to the effect of bias in medical research; bias is influenced by individual perspectives, money, and group think reinforced by the press. Studies are infrequently replicated and negative results are seldom reported. It is, therefore, critical that researchers acknowledge they are subject to these forces and work to overcome them. We each must expand our vision and not forget that our view of the world is affected by our personal filters that emerge from our experience and background. We must be vigilant and recognize our biases as we struggle to place the results of experiments in context and reveal the underlying mechanisms, that, for example in this case study, explain the reaction of beaked whales to sonar and determine the long-term effect of sonar on populations. Widening our filters will allow a variety of perspectives and backgrounds to inform management decisions.