Project Title: Humpback Whale Communication and the Effects of Anthropogenic Noise
Primary Investigator: Sean Hanser

Communication is important to many facets of organisms’ lives including food acquisition, reproduction, cultural transmission, and ecological relationships. It occurs via a variety of modalities such as vocal, chemical, and gestural communication, and the operating conditions for those modalities can be altered by “noise” in the environment. Noise that disrupts information exchange comes in many forms, and it may change the use of a communication system and possibly derail communication.
      Of all of their senses, cetaceans depend most upon their hearing to gather information about their environment and mediate their relationships with one another. Among the cetaceans, humpback whales are recognized by scientists and the public as one of the most vocally diverse and exciting species. Their most conspicuous vocalization has long been a focus of seamen and scientists, and is known as the "song" of the humpback whale. It is typically sung by male humpback whales during the breeding season. Although its function is not yet known for sure, many scientists think it is used to attract females and possibly signal other males to stay away from the singing male's area. It is unclear how much of humpback whales’ vocal communication system is represented by song, because the character of non-song vocalizations remains relatively unknown. Also, since song is predominantly a behavior of adult males, little is known about the vocalizations used by females and young whales. Song is a difficult place in which to start characterizing the humpback whale vocal repertoire. It has been dissected to some degree, but the classification of different components is somewhat arbitrary because the song is long, a continuous loop, and challenging to partition. Also, the form changes in structure and composition from year to year, making it a “moving target” for characterization.
      For this project, we would like to investigate humpback whale vocalizations other than song. The importance of general humpback whale communication is that it is used in social interactions and cooperative feeding behavior. It is fantastically diverse. For a small sample of humpback vocalizations, check out our sounds of cetacean research page.

RECORDING THE WHALES:
Over the years, Fred Sharpe has amassed a huge body of recordings from whales in Southeast Alaska, has been able to characterize the bubble net feeding behavior of the community of whales in Chatham Strait and Frederick Sound. When Sean Hanser, Brenda McCowan, and Laurance Doyle began our collaboration with AWF in 2004, we set out to take advantage of the rich resources AWF already had and to expand that resource with systematic, high-quality recordings.
      When we record in the field, we prefer to do so from smaller inflatable boats using a case full of recording equipment. We use high-quality underwater microphones, known as hydrophones, to capture the sound. Currently, our preferred hydrophone is made by Reson and has a nice broad response range: from 10 Hz, which is at the very low end of what the human ear can hear, up to 60 KHz, which is more than three times higher than an adult human can hear. Actually, we have trouble obtaining recording equipment that can record sounds that high, so our hydrophone has a greater range than our recording equipment. We need to use a preamplifier to boost the signal from the microphone before it goes into our recorder so we use a mixer, made by MACKIE, for that purpose. You can see the mixer clearly in the bottom left photo; it is the piece of equipment on the left that has many knobs on it. This is a more elaborate piece of equipment than we need right now, but when we eventually get to recording with an array of hydrophones, as we plan to do later, we will probably take better advantage of this piece of
equipment. We use a digital multitrack recorder made by Korg to capture the recordings and export them as sound files to our computer.
      The reason we use a multitrack recorder to gather recordings rather than a basic DAT recorder or something similar to that is that we need to make stereo recordings of several tracks at a time. While we are recording the underwater sounds, Sean is also recording a narration track, with a directional microphone, to make a log of what behavior and environmental information can be observed. This helps to set the context for recordings and explain what is occurring on a moment-by-moment basis.

Sean Hanser narrating during a recording of humpback whales foraging socially in Iyoukeen Cove, Chatham Strait.

IDENTIFYING INDIVIDUAL WHALES ACOUSTICALLY:
Through careful listening and sheer amount of time spent with the whales, Fred has managed to identify and classify, by ear, individual whales that produce the feeding call vocalization associated with bubble net feeding. Sean is taking Fred's recordings and analyzing them to see if there is a mathematical basis for discriminating among individual whales when they make the same vocalization, in this case the feeding call. We are expecting to be able to identify what acoustic features of the individual whales' vocalizayions make them individually identifiable.
      A statistical tool we are using to classify individual vocalizers is called Catgorization and Regression Trees, also known as CART. This process makes trees, called decision trees, using a variety of acoustic measurements we have made of our recordings to give us a sense of which measurements are the best for differentiating among individual whales. We have not yet been able to generate any nice clear trees for our whale vocalizations yet, but an example of decision tree that differentiates individual adult male Mexican spotted owls can be seen at the right. This tree was constructed for another research project that Sean and Brenda McCowan are participating in. You will notice that the tree only makes binary decisions at each juncture. This makes the sorting process clear and intuitive, but some trees can end up being quite extensive.
      For our whale recordings, very preliminary results suggest that the higher harmonics are important acoustic features for identifying individual whales. These are the higher-frequency components of sound that give it its color. They are not the main frequency you hear - that is called the fundamental frequency - but the other frequencies that are mathematically related to the fundamental frequency. This preliminary result is interesting, because we know that other mammals, including humans, use higher harmonics to differentiate among individuals of the same species.

THE EFFECTS OF ANTHROPOGENIC NOISE:
Many bioacoustic noise studies deal with high levels that have physical effects on the organism or that are loud enough to completely disrupt communication. Less has been done to investigate the potentially subtle effects of environmental noise that are equally disruptive. Subtle effects on communication can be undetectable to researchers if the organization of the communication system is not understood ahead of time. The mathematical discipline of information theory (IT) is used to assess the organizational complexity and optimal operating conditions of communication systems. Here, organizational complexity means how the components of a communication repertoire are arranged and combined to convey messages. IT has novel applicability in the field of conservation because it can be used to assess the effects of anthropogenic (human-generated) noise on the use of a communication systems by providing a quantifiable means of detecting and understanding departures from optimal communication. Marine mammals are particularly susceptible to negative impacts from anthropogenic noise. Their social interactions and feeding often depend largely on vocal communication. Humpback whales spends a large portion of the year within 10 miles of land, are frequently found in areas of high boat traffic, are social, and are therefore an ideal species for noise impact studies. If an effect from anthropogenic noise on humpback whales can be demonstrated, this can be used to help design management plans for conservation of whale habitat. Also, the novel process of dealing with noise at moderate levels by assessing departure from optimal communication would be a template for addressing other challenges of environmental degradation.
      Our previous research on bottlenose dolphin communication behavior demonstrated the applicability of IT by characterizing their whistle communication repertoire. As a result of that work, we are interested in applying IT to another marine mammal species with a complex vocal communication system for later comparison with bottlenose dolphins as well as demonstrating the applicability of IT to characterizing the impact of noise in the environment. There has been a great deal of study regarding the effects of noise on marine mammals particularly since vessels, military activity and research, industrial activity, and oceanic research can create tremendous amounts of noise. Results have been inconsistent, but anthropogenic noise has been speculated to cause disastrous results such as mass strandings and dulling of senses. Research in the area of noise has been dedicated primarily to understanding the hearing capabilities and physiology of marine mammals, what parts of the sound spectrum noise is generated in, and the physical force of loud sound in water.
      The collaborative group of UC Davis and AWF has been granted $49,100 by the National Fish and Wildlife Federation, to study the effects of vessel noise on humpback whale vocal behavior. Through systematic introduction of standardized moderate-level sound source experiments, we will assess the subtle effects of intrusive vessel noise on vocalization usage in humpback whales.