Physiological, morphological, and ecological tradeoffs influence vertical habitat use of deep-diving toothed-whales in the Bahamas

Dive capacity among toothed whales (suborder: Odontoceti) has been shown to generally increase with body mass in a relationship closely linked to the allometric scaling of metabolic rates. However, two odontocete species tagged in this study, the Blainville’s beaked whale Mesoplodon densirostris and the Cuvier’s beaked whale Ziphius cavirostris, confounded expectations of a simple allometric relationship, with exceptionally long (mean: 46.1 min & 65.4 min) and deep dives (mean: 1129 m & 1179 m), and comparatively small body masses (med.: 842.9 kg & 1556.7 kg). These two species also exhibited exceptionally long recovery periods between successive deep dives, or inter-deep-dive intervals (M. densirostris: med. 62 min; Z. cavirostris: med. 68 min). We examined competing hypotheses to explain observed patterns of vertical habitat use based on body mass, oxygen binding protein concentrations, and inter-deep-dive intervals in an assemblage of five sympatric toothed whales species in the Bahamas. Hypotheses were evaluated using dive data from satellite tags attached to the two beaked whales (M. densirostris, n = 12; Z. cavirostris, n = 7), as well as melon-headed whales Peponocephala electra (n = 13), short-finned pilot whales Globicephala macrorhynchus (n = 15), and sperm whales Physeter macrocephalus (n = 27). Body mass and myoglobin concentration together explained only 36% of the variance in maximum dive durations. The inclusion of inter-deep-dive intervals, substantially improved model fits (R2 = 0.92). This finding supported a hypothesis that beaked whales extend foraging dives by exceeding aerobic dive limits, with the extension of inter-deep-dive intervals corresponding to metabolism of accumulated lactic acid. This inference points to intriguing tradeoffs between body size, access to prey in different depth strata, and time allocation within dive cycles. These tradeoffs and resulting differences in habitat use have important implications for spatial distribution patterns, and relative vulnerabilities to anthropogenic impacts.

The body masses of our study species were reported inconsistently across the literature, using a variety of different summary statistics (e.g., maximum, mean, approximation). To standardize estimates of body mass used in the allometric analyses reported in this paper we developed estimates of median body masses in the different sexes and age classes of our study species using standard length and body mass measurements from strandings and historic whaling records from around the world. We also collected myoglobin concentration ([Mb]) measurements from a variety of published sources for consideration in the subsequent modeling of dive durations and dive depths.

Methods
Standard length and body mass measurements were retrieved from the National Museum of Natural History (NMNH) whale collection database (http://collections.nmnh.si.edu/search/mammals/ accessed on July 13, 2015), as well as a variety of primary literature sources not included in the NMNH database [1][2][3][4][5]. The mass data provided by all of these sources was relatively sparse and systematically biased towards individuals with smaller standard lengths in many species, likely due to the logistical challenges of measuring mass in animals at the scale of many cetaceans. This systematic bias presented challenges in estimating representative median masses for each species, sex, and age class based on the mass measurements alone. To overcome this challenge, we estimated power-law relationships of body mass to standard length for each species. Following the approach of Kasuya and Matsui (1984) [4], this relationship was modeled as a linear function of log-transformed body mass and standard length measurements using lm function from the R library stats. Extensive standard length measurement data from NMNH and other sources [2,3] were then used to estimate the median standard length of each species, sex, and age class and these values were then used to predict the median mass of each species, sex, and age class from the power-law relationships detailed above. Threshold lengths separating sub-adult from adult age classes in the standard length dataset were defined based on reported values [2,3] and on the minimum length at which sexual maturity characteristics, such as evidence of pregnancies and increases in gonadal mass, were described in necropsy notes.

Results
Across the five tagged species, we compiled 175 simultaneous measurements of mass and standard length from a variety literature and unpublished sources from around the world. We compared these data with 1,912 records in which standard length alone was measured. Estimates of median body mass from power-law models relating body mass and standard length yielded adult body masses spanning nearly two orders of magnitude from P. electra (169.6 kg) to P. macrocephalus (11774.5 kg). Adult P. macrocephalus typically exhibited extreme sexual size dimorphism (male:female body mass ratio of 2.3), however no adult males were tagged in this study. In the modeling of dive duration and dive depth, sub-adult male body mass estimates (median 12362 kg) were used, which were only slightly greater than the body mass estimates of adult females (median 11186 kg; Table 2).
In three of our five study species mean [Mb] concentration values were based on measurements taken across multiple individuals (Ziphius cavirostris, n = 2, Mesoplodon densirostris, n = 2, Globicephala macrorhynchus, n = 6). Among these measurement derived from Noren and Williams et al. (2000) [7], and Velten et al. (2013) [10], the reported standard deviation values represented <10% of mean values, suggesting that inter-individual variation was present but relatively small. The [Mb] value for P. electra predicted from eq. 1 in Mirceta et al. (2013) [9], was substantially lower than the empirical mean [Mb] concentration value measured in the closely related G. macrorhynchus (Table A).

Table A. Summary of standard-length (m), mass (kg), and myoglobin (mg kg -1 ) measurements in five species of odontocete cetaceans.
Median mass was estimated from the models presented in Fig A in S1 Appendix. Threshold standard-lengths were used to categorize measured individuals into age classes within each sex and species. The sources of   The histograms embedded in the upper margin of each plot indicate the distributions of male (lighter shade, partially transparent) and female (darker shade) standard lengths recovered from stranding and historic whaling data. The median adult standard lengths for each species were used to estimate the median adult body masses for each species and sex using the fitted power law relationship indicated in each plot. Subadult male body mass was also estimated for P. macrocephalus.