Marine Mammals: The Smithsonian Magazine Talks About NOC

Disclosure: I am a researcher who worked with NOC at the US Navy Marine Mammal Program.

Smithsonian Magazine ran an article by Charles Siebert titled, “The Story of One Whale Who Tried to Bridge the Linguistic Divide Between Animals and Humans”. This was ostensibly about the work Dr. Sam Ridgway and others at the US Navy Marine Mammal Program did to study vocal mimicry in a white whale named NOC (pronounced “no-see”).

The article splits over two pages. Page one is mostly just discussion of events. Page two, though, is an agenda-driven mess that runs off the rails. Here’s something Siebert deems worthy of inclusion in the Smithsonian Magazine:

Samantha Berg, who now runs an acupuncture clinic with her husband in Palmer, Alaska, vividly remembers the time she heard the recording of Noc.

“It gave me chills,” she said. “There’s something about the quality of the sound that’s deeply disturbed. Like if you woke up at 3 o’clock in the morning and heard somebody screaming in distress. Whether such a sound comes from another human or not, I think there is some part of us that is able to recognize distress. I’ve heard a lot of beluga noises, and I’ve never heard a beluga sound like that.”

This is not one of Siebert’s finer moments. Berg is active in the animal rights movement, something Siebert fails to inform his readers about, while taking up space to give them the irrelevancy about acupuncture. Berg doesn’t have research experience with white whale vocal mimicry, and gives us straight, undisguised anthropomorphism as her insight. And I don’t mean simply that Berg is attributing human characteristics to white whales, but rather that Berg’s tuning of her internal distress meter is likely also bound to human characteristics. Voice stress analysis relies upon the physics of human sound production in the larynx and its vibrating “cords”, where stress tends to tighten musculature and has an impact on the fundamental frequencies used in human speech. That’s all very well for off-hand pronouncements on recordings of humans, but Berg isn’t quoted as to how she modulates her assessment for the fact that NOC isn’t producing those sounds with vocal cords in a larynx at all. Clicks and whistles in the white whales get produced at the phonic lips, paired structures in the nasal passages dorsal to the bony nares and ventral of the blowhole. NOC is being a vocal mimic via a different sound-producing mechanism. And yet here we have a confidently-stated opinion about the internal mental state of a member of a different species, based on zero personal interaction with the subject, and just listening to a recording of sounds produced in ways no human can manage at all. (We don’t have phonic lips in our nasal passages.) Nor does Siebert offer the reader any guidance on the caveats that inevitably must attend an unsubstantiated assertion of the sort he passes along from Berg.

Siebert delves into bioacoustics:

Once inside SeaWorld San Diego’s front gates, one is immediately subsumed by sound. Calypso Muzak clashes with rock ’n’ roll and these with the blaring triumphal trumpets and 60-foot-boat-slide screams from the popular water ride, Journey to Atlantis. Loudspeakers issue repeated reminders of the fast dwindling seats for upcoming shows at the Shamu and Dolphin stadiums. It’s impossible to imagine what all that sounds like to whales: creatures who can hear one another across oceans.

Empirical research can succeed where imagination fails. I think it would be a good thing to actually make the measurements that would allow researchers to talk sensibly about the acoustic experience of captive marine mammals. But the blithe assertion that it must be horrible because of loudspeakers simply isn’t substantiated by anything other the the author’s anthropomorphism, and the further implication that we can’t possibly come to know anything more about it is simply untrue. We do know how to do this. This would involve a calibrated microphone and hydrophone. We would also need in-air and underwater audiograms of each of the species (and often the individuals) we are interested in. One principle in acoustics is that interfaces with a high impedance mismatch, like those of air and water, do not efficiently transmit sound energy. So we already know that most of the in-air sound energy Siebert discusses is going to stay in air, and very little of that will make it underwater. Don’t take my word for it. This can be confirmed with the calibrated microphone and hydrophone by recording a sound generated in air and see that it is much reduced underwater. Also, marine mammals whose in-air audiograms have been taken typically show reduced sensitivity to sound in air, so our expectation from actual measurements is that the in-air sound is not going to be perceived as being as loud to those marine mammals as it is in we humans, because they don’t have great sensitivity to the sound in air, where it is loudest, and also because in water, where their hearing is acute, little of the energy of the in-air sound was able to make the transition.

It is true that the physics of underwater sound make it possible for certain species to emit calls that span ocean basins or significant parts of them. What Siebert doesn’t find useful to pass on to his readers is that where that happens, it involves very low frequency sound, and only a few species of whale are known to produce it. The physics of underwater sound means that higher-frequency sound is attenuated relatively quickly, primarily via MgSO4 relaxation. Dolphin echolocation range, for example, only tests out to a bit over 100 meters for a 7.5cm target (IIRC). That’s still pretty impressive, but nowhere close to the “span ocean basins” claim. So which whales are perhaps sharing sound over ocean basins? Those would be various large baleen whales, whose calls often fall into the category of infrasound, with peak frequencies at or less than 20 cycles per second. Beluga whales are toothed whales, and are not in that group. Neither are dolphins. None of the large baleen whales are kept long-term in captivity here in our parks. (Sea World San Diego did rescue, rehabilitate, and release two gray whale calves a couple of decades apart.) So it would appear that Siebert is engaging in hyperbole in his final assertion in the quote above, though it is hyperbole that his readers, judging by the comments, are not likely to recognize as such.

Siebert then touches upon work I was directly involved with, but I only recognized this because Siebert used the name of the project, “Deep Hear”. (Among other things, I designed the project T-shirt way back in 1995.)

Four years after Noc launched into his talking spree, he just as abruptly stopped, reverting in 1999 to “Beluga” for the rest of his days in captivity. In the early 1990s, with the end of the cold war and cutbacks in defense spending, the Marine Mammal Program was greatly downsized. By then only three belugas remained in the Point Loma enclosure, Muk Tuk, Noc and Ruby. Lyl had succumbed to pneumonia only two years into his training. Chr died of a lung infection in 1984. Churchill died in 1987 of pneumonia. In April 1997, Ruby was transferred to SeaWorld, though she remains the property of the U.S. Navy.

Muk Tuk and Noc, however, were retained for a new operation, dubbed Deep Hear. It was developed to test the potential effects on marine mammals of a new type of low-frequency active sonar (LFAS), which has since been cited by more than 100 scientists as the cause of massive whale strandings. The noise often induces entire pods to surface so rapidly in an attempt to escape it, they die of a condition to which scientists had assumed whales to be immune: the bends. In 2001, Muk Tuk joined Ruby behind the glass at Wild Arctic. She died of a lung infection in 2007.

There is so much error and confusion in there that it is problematic where to start on corrections. As good a place to start is the first quoted sentence there, the one that claims that NOC was observed to abruptly stop his vocal mimicry and “revert” to standard white whale vocal behavior for some indefinite period of time, starting in 1999. There just isn’t any there there. Siebert himself had already noted a significant event in NOC’s life in 1999: he died.

He ended up being deployed for two top-secret Navy surveillance and retrieval programs before succumbing to meningitis in 1999 at the age of 23 while still in the Navy’s care.

Dying would be a pretty non-mysterious way to stop talking, all right. This is an indication that Siebert has seriously mangled the chronology. Siebert’s sentence construction would lead the reader to infer that the Deep Hear project was initiated on or after 1997 at least, and perhaps 1999. In fact, a commenter made exactly that inference:

Lorika13

Anyone else catch that Noc mysteriously reverting back to Beluga only speech coincided with the new mission with the low-frequency active sonar testing? It was pretty well glossed over in the story, but I bet there’s a correlation there. If the sound causes whole pods to commit suicide, just imagine what it would feel like to him when he’d been trying to communicate with us. Or, did the frequencies even damage his brain? Maybe he couldn’t speak human anymore after being subjected to those frequencies. It’s all fascinating, but a sad story all the same.
[…]

But the way Siebert constructed his text makes this confusion understandable, if regrettable. Let’s see what the original paper says about NOC’s spontaneous vocal mimicry behavior:

The speech-like behavior subsided after about four years. After the whale matured, we no longer heard speech-like sounds. However, NOC remained quite vocal. He produced typical echolocation pulses with peak frequencies between 60 and 120 kHz, whistles with fundamental frequencies of 2 to 10 kHz and various pulse burst sounds previously described as “squawks, rasps, yelps or barks”.

Given that the paper notes 1984 as when NOC was identified as the source of vocal mimicry, that would make 1988 or perhaps 1989 the correct timeframe for what the paper’s authors discussed. It seems likely that Siebert intended to write 1989, but ended up with 1999 instead. Notice, too, that the paper makes no reference to either an “abrupt” transition in behavior (and, in fact, standard English usage of “subsided” would argue against that) or that NOC’s change in vocal behavior constituted a reversion instead of a maturation.

While we are still on the topic of chronology, data collection for Deep Hear was completed in 1995.

There is a long-running confusion, amounting to obsession, in the animal rights community concerning “low-frequency sonar”. Siebert carries on in this dubious tradition, casting discussion of the Deep Hear project in terms of effects of low-frequency sonar. This error also misled “Lorika13”, as seen in the quoted comment. There is a naval sonar technology that is unambiguously linked to damage and mortality in cetaceans, with beaked whales being the most susceptible, but that technology is mid-frequency sonar. I was part of the 2002 workshop looking into the “tissue bubble” damage hypothesis proposed as a mechanism of morbidity and mortality of mid-frequency sonar systems; our report is freely available.

The notion hinted at by Siebert that NOC and MUK were exposed to low-frequency sonar signals as part of the Deep Hear experiment is not just wrong, it is absolutely contradicted by the research paper reporting our methods and results (which is, hey looky here, also freely available online). If you look in the methods, you will find that the sounds presented to the whales were a trainer’s projected “bridge” stimulus and 0.5 second pure tones produced by a function generator (look for “Wavetek 275” in the report). There wasn’t any use of a naval sonar system in the experiment, low-frequency or otherwise. How much effort would it have taken to look at the actual source materials and accurately relay them to readers? The point of the Deep Hear study was aimed at finding the lowest-amplitude sounds the whales could hear, not to expose them to high-amplitude noise. Exposing them to high-amplitude noise would, if anything, make it harder to establish the hearing thresholds that were the whole point of the project, which anyone actually reading the freely available report ought to understand. Why didn’t Siebert understand this, or inform his readers about it? The methods of Deep Hear in studying hearing are all standard psychophysics techniques that you yourself might experience in your next trip to your audiologist, except that you won’t be asked to dive to do them.

The Deep Hear study had non-mysterious results, too, though Siebert apparently did not find them interesting enough to pass on to his readers: the hearing of whales remains just as sensitive at depth, and thus policy on noise mitigation cannot rely on the notion that mammalian hearing sensitivity decreases with depth. That was known to be true for humans and chinchillas, but Deep Hear demonstrated that it was not true for white whales, and thus unlikely to be true for the rest of the cetacea. This is in many ways an inconvenient result for both commercial and military interests, and yet there was no interference in our ability to publish these findings.