Search for a command to run...
Body size is a fundamental feature of an individual animal's physiological and life history state (Christiansen et al. 2018; Pirotta et al. 2024; Stewart et al. 2021; van Aswegen et al. 2025) but can be difficult to quantify in large marine mammal taxa. Body size can be proxied by quantifying dominant stroke cycle frequency (Isojunno et al. 2018), which has been shown to correlate with body size across species (Sato et al. 2007), but opportunities for validation of this effect within a single species are limited. Here, we use a unique dataset from contemporaneous tag deployments and UAV (drone) photogrammetry overflights to directly validate dominant stroking frequency as a proxy for total body length in free-ranging sperm whales (Physeter macrocephalus) and parameterize functions for converting between dominant stroking frequency and body length, volume, and mass. Data were collected off the Azores archipelago, Portugal during June–September 2021 and 2022, and off Lofoten, Norway in August–September 2019. Tagging and drone flights in the Azores were undertaken under research permits LMAS-DRAM/2020/06 and 2021/12, issued by the Azores Regional Directorate for Sea Affairs (Direção Regional dos Assuntos do Mar; DRAM) and held by coauthor MAS. Off Lofoten, the same activities were undertaken under a permit issued by the Norwegian Animal Research Authority (Permit #18-126201). All responses to UAV overflights and tagging approaches were scored on a scale of 0–3 according to standard behavioral response criteria (Weinrich et al. 1992). If a response of type 3 (severe and prolonged) was observed, research activities with that individual were immediately ceased. All research activities were approved by the Animal Welfare and Ethics Committee of the University of St Andrews, UK. The suction-cup attached tags contained a DTag 3 sensor package recording sound and movement (Johnson and Tyack 2003). Tags were applied from small boats using either a 12 m cantilevered pole or a 7 m hand-pole. A DJI Phantom 4 Pro UAV was used for all flights performed in the Azores in 2021–2022, whereas a DJI Phantom 4 was used for flights in Norway. Overhead photogrammetry images of sperm whales were obtained by flying the UAV above the animals at an elevation of 13–23 m in winds not exceeding 6 m s−1 and sea state not exceeding Beaufort 3. Images were scaled using altimetry data from a downward-facing laser range-finder (a Lightware Laser SF11C [LightWare Optoelectronics. 2018], driven by a data logger custom manufactured by Kelp Marine Research [Kelp Marine 2024]) and mounted on the UAV as additional payload. All photogrammetry images were obtained during the corresponding tag deployment (for more detailed descriptions of the fieldwork, see Burslem 2024; Burslem et al. 2025a). We identified the dominant stroking signal, following Sato et al. (2007). We marked the periods of ascent, during which sperm whales exhibited more consistent and stereotyped steady swimming than during descent or bottom phases, and identified periods of ascent during which there was regular stroking with stable pitch and roll signals. We then identified the dominant stroking frequency during these periods manually by inspecting spectral density plots of the heave and surge vector acceleration signals and marking the peak associated with stroking (Sato et al. 2007). Whale total length (L) and width at 5% intervals along the body were measured using Morphometrix software (Torres and Bierlich 2020). Volumes were estimated from these measurements following the methods described by (Christiansen et al. 2019) and using the body site-specific height: width ratios estimated for sperm whales from a combined sample of adults and calves by Glarou et al. (2023). The volume of each segment (Vs; the region between adjacent measurement sites) was modeled as a series of infinitesimal ellipses (Christiansen et al. 2019; their equation 1). Segment volumes were then summed to give the estimated total body volume (Christiansen et al. 2019; their equation 2). Body volume can then be converted to mass using published estimates of total body density. Statistical analysis consisted of modeling dominant stroking frequency as a function of total body length to test the hypothesis that larger individuals used a lower dominant stroking frequency to estimate the model parameters and to assess the confidence of any resulting predictions. This model was fitted as a (natural) log–log standardized major axis model with robust estimation using the sma function from smatr software (Warton et al. 2012). For the purposes of prediction, and to facilitate comparison with other studies, we fitted an additional function to predict stroking frequency as a function of body volume. Of 17 animals for which both tag and overhead photogrammetry data were collected, 15 performed enough steady ascent swimming to estimate dominant stroking frequency. Log–log linear modeling showed strong statistical support for a negative relationship between body length (range 8.4–15.2 m, Table 1) and dominant stroking frequency f s $$ {f}_s $$ with the (response scale) form: f s = 12.16 · L − 1.76 $$ {f}_s=12.16\cdotp {L}^{-1.76} $$ (log-scale estimates and 95% CI ranges: intercept = 2.50 [1.11, 3.89], slope = −1.76 [−2.45, −1.26], R2 = 0.82, p < 10−5, n = 15, Figure 1A,B). The corresponding fitted model of dominant stroking frequency as a function of body volume was f s = 1.06 · V − 0.61 $$ {f}_s=1.06\cdotp {V}^{-0.61} $$ (log-scale intercept = 0.060 [−0.59–0.71], slope = −0.61 [−0.89, −0.42], R2 = 0.83, p < 10−5, Figure 1C,D). The two models had almost identical fit and forward model selection did not support the inclusion of volume once length was known, or vice-versa (AIC of f s $$ {f}_s $$ ~ volume, f s $$ {f}_s $$ ~ length and f s $$ {f}_s $$ ~ volume + length models = −13.3, −12.6, and −11.9, respectively). For the purpose of comparison with the results of Sato et al. (2007), we converted volume to mass assuming a tissue density of 1029.7 kg m–3 adjusted for air spaces (i.e., lungs and acoustic apparatus) of 25.6 mL kg−1, estimated from 48 tagged sperm whales, resulting in a whole body density of 1003.3 kg m−3 (Burslem et al. 2025a & equations 10–13 therein, 2025b). The allometric relationship estimated here corresponds with the cross-species relationship reported by Sato et al. (2007) of: f s = 3.56 · M − 0.29 $$ {f}_s=3.56\cdotp {M}^{-0.29} $$ , but not perfectly (see Figure 2 for a comparison of the two models over the range of body masses estimated in this study). A clear difference in kinematics between sperm whales and the other species studied by Sato et al. (2007) is that, due to their extremely long skulls, a relatively smaller proportion of large sperm whales' bodies is articulated compared with other species. While overall, sperm whales gross body geometry remains relatively constant across size (volume allometry scaling exponent = 2.9; Burslem et al. 2025a; Glarou et al. 2023), their skulls and acoustic apparatus get proportionately larger with size, decreasing the articulated proportion of the body as sperm whales get larger. However, this would be expected to lead to a smaller estimated rate of change in stroking frequency per unit total length rather than a larger one, as found here (Figure 2). It seems more likely, therefore, that the differences were driven either by the small sample size in this study (n = 15), or the broad range of propulsion modes and body geometries of the species included in Sato et al.'s (2007) cross-species analysis. Comparing the sperm whale model to a cross-species one fitted only to geometrically similar cetaceans, as Sato et al. (2010) did in penguins, could help to elucidate the additional drivers of stroke cycle frequency while controlling for propulsion mode and geometry. The analysis of body length and dominant stroke cycle frequency demonstrated the expected negative correlation with high confidence (Figure 1). This result confirms that, although likely to be noisy compared to photogrammetric measurements, a body size signal is detectable in the dominant stroking frequency in sperm whales, and supports the use of this method as a relative proxy for cetacean body size (Isojunno et al. 2018). The allometric equations presented here allow conversion of dominant stroke frequencies into approximate length (in m) or volume (in m3) in sperm whales using the forms: L = 0.0822 · f s − 0.568 $$ L={\left(0.0822\cdotp {f}_s\right)}^{-0.568} $$ and V = 0.942 · f s − 1.646 $$ V={\left(0.942\cdotp {f}_s\right)}^{-1.646} $$ , respectively. We envision that these functional forms could be used to estimate the body size of sperm whales in real units from existing and future tag records. Alec Burslem: conceptualization, methodology, software, data curation, investigation, formal analysis, visualization, project administration, resources, writing – original draft, writing – review and editing. Saana Isojunno: conceptualization, methodology, software, data curation, investigation, formal analysis, supervision, funding acquisition, visualization, project administration, resources, writing – review and editing. Rui Prieto: methodology, investigation, project administration, resources, writing – review and editing. Mónica A. Silva: methodology, investigation, project administration, resources, writing – review and editing, funding acquisition. Patrick J. O. Miller: funding acquisition, supervision, formal analysis, investigation, data curation, software, methodology, conceptualization, visualization, writing – review and editing, resources, project administration. This work was supported by The United States Office of Naval Research (ONR; 00014-17-1-2757), FCT - Fundação para a Ciência e Tecnologia, I.P. by project references UID/05634/2025 and UID/PRR/05634/2025 (DOI: 10.54499/UID/05634/2025; 10.54499/UID/PRR/05634/2025), and from the Regional Directorate for Science, Innovation and Development, through the PROSCIENTIA Incentive System under project M1.1.A/FUNC.UI&D/014/2025. Data collection in the Azores was also supported by Fundo Regional da Ciência e Tecnologia (FRCT), FCT, and the EU through projects WATCH IT (Acores-01-0145-FEDER-000057), FCT-Exploratory (IF/00943/2013/CP1199/CT0001), META (FA_06_2017_017) and SUMMER (H2020, GA 817806) co-funded by FEDER, COMPETE, QREN, POPH, ESF, PO AZORES 2020, the Portuguese Ministry for Science and Education. Author AB was supported in part by a PhD studentship from the University of St Andrews School of Biology. The authors declare no conflicts of interest. All underlying data are reported in the paper.