Kathryn A. Cortopassi

THE COMPARISON OF HARMONICALLY RICH SOUNDS USING SPECTROGRAPHIC CROSS-CORRELATION AND PRINCIPAL COORDINATES ANALYSIS

ABSTRACT We explore the effectiveness of spectrographs cross-correlation (SPCC) combined with principal coordinates (PCO) analysis as a method for sound comparison. We do this using synthetic sounds modeled after the individually-distinctive, harmonically-rich contact calls of wild orange-fronted conures Aratinga canicularis. Calls with acoustic properties similar to Aratinga contact calls are common in other taxa including non-oscine birds, primates and cetaceans. We generated signals with known variations in time-frequency pattern, duration, noise level, harmonic content and harmonic weighting, and applied SPCC-PCO analysis to obtain an ordering of sounds in n-dimensional space. We find that shared time-frequency patterns dominate the positioning of sounds in PCO space. This was true despite high variability in signal-to-noise ratio (from −60 to +40 dB) and duration (150–275 ms). Furthermore, inclusion of naturally-weighted harmonics (versus fundamentals only) enhances, rather than obscures, the separation of call types. We conclude that SPCC-PCO is an effective method for sorting sounds based on overall time-frequency pattern. In addition, the resulting PCO measures can be used in statistical tests of association with extrinsic variables. The method is thus an effective starting point for examining most bioacoustic hypotheses.

GEOGRAPHICAL VARIATION IN THE CONTACT CALLS OF ORANGE-FRONTED PARAKEETS

Abstract We examined patterns of geographic variation in contact calls of Orange-fronted Parakeets (Aratinga canicularis) during the nonbreeding season. Calls were recorded throughout the range of that species in Costa Rica. Recordings of wild-caught birds held for one to two weeks indicated that each individual favors one dominant call type and different birds use different favored calls. We used that fact to cull replicate calls from the same individual in field recordings of flocks observed at various locations throughout the sample area. Remaining recordings from a given year were submitted to spectrographic cross-correlation and principal coordinates analysis as described by Cortopassi and Bradbury (2000). Principal coordinates were then correlated with site location using MANOVA. Call durations were also examined for geographical patterns. Like sympatric Yellow-naped Parrots (Amazona auropalliata) studied over the same region by Wright (1996), Orange-fronted Parakeets exhibited significant geographic variation but, unlike the larger species, showed no discrete dialect patterns. Call duration varied clinally but with different patterns for the Nicoya Peninsula and the Guanacaste mainland. Two principal coordinates also showed clinal variation even after removing any correlated duration effects. Scale over which local calls were statistically homogeneous was 7–10 km. We compared that figure to home ranges of 18 birds radio-tracked concurrently with call sampling. Both range areas and range lengths were asymptotic after a week of tracking. Asymptotic range lengths were 2–9 km. Scale of movements of birds, at least during that period, was thus similar to distance within which calls tended to be statistically homogeneous. This study and that by Wright show that several well-known patterns of geographical variation seen with passerine song are replicated quite closely in contact calls of parrots, despite the fact that the functions of vocal signals are quite different in the two taxa.

Vertical transmission of learned signatures in a wild parrot

Learned birdsong is a widely used animal model for understanding the acquisition of human speech. Male songbirds often learn songs from adult males during sensitive periods early in life, and sing to attract mates and defend territories. In presumably all of the 350+ parrot species, individuals of both sexes commonly learn vocal signals throughout life to satisfy a wide variety of social functions. Despite intriguing parallels with humans, there have been no experimental studies demonstrating learned vocal production in wild parrots. We studied contact call learning in video-rigged nests of a well-known marked population of green-rumped parrotlets (Forpus passerinus) in Venezuela. Both sexes of naive nestlings developed individually unique contact calls in the nest, and we demonstrate experimentally that signature attributes are learned from both primary care-givers. This represents the first experimental evidence for the mechanisms underlying the transmission of a socially acquired trait in a wild parrot population.

Hearing and Vocalizations in the Orange-Fronted Conure (Aratinga canicularis)

The auditory sensitivities of the orange-fronted conure (Aratinga canicularis) were examined in relation to the spectral characteristics of its vocalizations. Absolute thresholds, masked thresholds, frequency difference limens, and intensity difference limens for pure tones were obtained using psychoacoustic techniques. In general, hearing abilities are similar to those found in many avian auditory generalists. One exception is the unusually low critical ratio (masked threshold) between 2.0 and 4.0 kHz, similar to that previously found in the budgerigar (Melopsittacus undulatus). These auditory sensitivities were compared with average spectra for (a) contact calls and (b) a general sample of vocalizations recorded from wild birds. The spectral regions of both greatest vocal energy and best auditory sensitivity were between 2.0 and 5.0 kHz.

Characteristics of gunshot sound displays by North Atlantic right whales in the Bay of Fundy

North Atlantic right whales (Eubalaena glacialis) produce a loud, broadband signal referred to as the gunshot sound. These distinctive sounds may be suitable for passive acoustic monitoring and detection of right whales; however, little is known about the prevalence of these sounds in important right whale habitats, such as the Bay of Fundy. This study investigates the timing and distribution of gunshot sound production on the summer feeding grounds using an array of five marine acoustic recording units deployed in the Bay of Fundy, Canada in mid-summer 2004 and 2005. Gunshot sounds were common, detected on 37 of 38 recording days. Stereotyped gunshot bouts averaged 1.5 h, with some bouts exceeding 7 h in duration with up to seven individuals producing gunshots at any one time. Bouts were more commonly detected in the late afternoon and evening than during the morning hours. Locations of gunshots in bouts indicated that whales producing the sounds were either stationary or showed directional travel, suggesting gunshots have different communication functions depending on behavioral context. These results indicate that gunshots are a common right whale sound produced during the summer months and are an important component in the acoustic communication system of this endangered species.

Automatic Animal Tracking Using Matched Filters and Time Difference of Arrival

A method for tracking animals using a terrestrial system similar to GPS is presented. This system enables simultaneous tracking of thousands of animals with transmitters that are lighter, longer lasting, more accurate and cheaper than other automatic positioning tags. The technical details of this system are discussed and the results of a prototype are shown.

Hearing and vocalizations in the orange‐fronted conure (Aratingacanicularis), a small parrot

The orange‐fronted conure (Aratinga canicularis) is a small Neotropical parrot that is the subject of an ongoing study of vocal communication and social organization. Three captive‐bred orange‐fronted conures were trained to detect pure tones using operant conditioning with a food reward and tested using the method of constant stimuli. These birds were tested at seven frequencies ranging from 0.5 to 8 kHz. The greatest sensitivity was found between 1 and 4 kHz and lower sensitivity outside this region. The shape of the audiogram for this species was similar to those previously found for three other species of small parrots, but the absolute sensitivity for orange‐fronted conures was generally lower than these other parrot species at most frequencies. The range of greatest sensitivity in the orange‐fronted conures corresponded fairly well with the spectral characteristics of their contact calls. In a sample of 288 contact calls recorded from eight wild birds in Costa Rica, the mean peak frequency of the calls was 3371 Hz with a SD of 341 Hz, and the bulk of spectral energy fell between 2295 and 5231 Hz. [Work supported by NIH Grant No. R01 DC00198 (RJD), NRSA Fellowship Nos. DC00046 and MH12111 (TFW), and a NSF grant (JB and KAP).]

Seasonal ambient noise levels and impacts on communication in the North Atlantic right whale

The North Atlantic right whale is a highly endangered species of baleen whale. Acoustic communication plays an important role in the social behavior of these whales. Right whales are found in coastal waters along the east coast of the United States, an area characterized by high levels of human activity. Most of these activities generate noise that is propagated into the coastal marine environment. The goals of this project are to characterize the noise, both natural and anthropogenic, in right whale habitat areas to determine what levels of noise the whales are regularly exposed to, and whether the acoustic behavior of right whales changes in response to increased noise. Continuous recordings were made from autonomous bottom‐mounted recorders in three major habitat areas in 2004 and 2005; Cape Cod Bay (December–May), Great South Channel (May), and the Bay of Fundy, Canada (August) to passively detect right whales by recording their vocalizations. Here, we describe the ambient noise levels in these record...

Detection of FM signals in the presence of non‐Gaussian noise

In bioacoustics, the problem of detecting frequency‐modulated signals in the presence of non‐Gaussian noise is of great interest. Matched filters (MF) are often chosen over frequency‐specific energy (FSE) detectors because of their improvement in both signal gain and target specificity. Under non‐Gaussian conditions however, MF detectors do not ensure acceptable trade‐off between false alarm and missed detection rate. To decrease false alarm rate, we propose a two‐stage detection technique. First, a MF is applied, with threshold prescribed by the acceptable false alarm rate, to generate candidate detections. Second, a signal recognition (SR) algorithm is applied to the candidates. The SR algorithm estimates modulation parameters from the signal spectrogram, and a detection decision is made based on how well parameters match a criterion set for the signal of interest. We applied this technique to right whale contact calls recorded in Cape Cod Bay. Results demonstrate that the MF‐SR technique decreases fals...