An analysis of pilot whale vocalization activity using hidden Markov models

Abstract

Vocalizations of cetaceans form a key component of their social interactions. Such vocalization activity is driven by the behavioral states of the whales, which are not directly observable, so that latent-state models are natural candidates for modeling empirical data on vocalizations. In this paper, we use hidden Markov models to analyze calling activity of long-finned pilot whales (Globicephala melas) recorded over three years in the Vestfjord basin off Lofoten, Norway. Baseline models are used to motivate the use of three states, while more complex models are fit to study the influence of covariates on the state-switching dynamics. Our analysis demonstrates the potential usefulness of hidden Markov models in concisely yet accurately describing the stochastic patterns found in animal communication data, thereby providing a framework for drawing meaningful biological inference.

Individual, ecological, and anthropogenic influences on activity budgets of long‐finned pilot whales

“…We showed that this approach can be used to generate ethograms from animal-borne tag data, which is a key challenge to describe the behavior of free-ranging species (Sakamoto et al 2009). Such an approach could also be used to model, for example, the context specificity of vocalizations (Popov et al 2017), a subject of interest in social communication. Hidden Markov models account for the time series nature of the data explicitly, but a potential drawback is the simplistic first-order Markov aspect of the model that assumes that the underlying state probability is only dependent upon the previous state.…”

Penalized estimation of flexible hidden Markov models for time series of counts

“…Le Strat & Carrat, 1999), ecology (pilot whale vocalization counts; Popov et al, 2017), and bioinformatics (T-lymphocyte counts; Marino et al, 2018).…”

Study of the relationship between pilot whale (Globicephala melas) behaviour and the ambiguity function of its sounds

“…1): clicks, whistles and pulsed calls 3,9,10 . Clicks are rapid and, usually, repetitive bursts of short, broadband sounds, primarily used for echolocation 1,11,12 ; whistles are continuous tonal sounds with few or no harmonics, possibly used for maintaining the cohesion of the group during foraging or travel 10,13,14 ; pulsed calls are rapidly broadband sound pulses, with distinct tonal properties caused by high pulse-repetition rates, possibly used for communication 3,10 . To obtain these conclusions, much significant work has been done: Kellogg et al 1 proved that clicks are used for echolocation by analysing the time-domain waveform of odontocetes' sound; Au et al 11 analysed the frequency spectra and source levels of clicks to prove that clicks can be used for echolocation; Fais et al 12 found that sperm whales use high-frequency, low amplitude clicks to provide high-resolution biosonar updates during the last stages of capture; Nemiroff 3 , Ford 10 , and Weilgart and Whitehead 13 obtained the conclusion that whistles and pulsed calls can be used for communication by analysing the time-frequency spectra of odontocetes sound; Popov et al 14 While the pilot whale sound was being recorded, the pilot whale kept swimming near the ship.…”

“…The pilot whale may use sound for a variety of purposes such as detecting partners or prey, communicating with each other etc.. However, it is very difficult to identify the specific purpose of a certain type of sound only by recording the sound and observing the swimming of the pilot whale 14,15 .…”

“…As demonstrated already in literature 16,17,18 , the wideband ambiguity function is an effective tool to evaluate the five parameters of sound waveform. Therefore, different from some traditional analysis methods 1,3,10,11,13,14,19 , in this paper, we use the wideband ambiguity function 16,18 to analyse different types of sound produced by the pilot whale. Through statistical analysis of the wideband ambiguity functions for a large collection of different sounds (clicks, whistles and pulsed calls), we find that the clicks have the best range resolution and their range ambiguity functions have the lowest sidelobe-to-mainlobe suppression ratio; pulsed calls and one part of whistles have excellent speed resolution and small Doppler tolerance and their speed ambiguity functions have the lowest sidelobe-to-mainlobe suppression ratio; the other part of whistles has large Doppler tolerance.…”