Using 3D models to improve estimates of marine mammal size and external morphology

Abstract

Studies of animal morphology inform our understanding of many combined aspects of biology, including thermal ecology, biomechanics and energetics. Studies that accurately describe the size and shape of marine mammals, for example, can be particularly useful in understanding the challenges an endotherm faces when moving and thermoregulating in an aquatic environment. The relationship between surface area (SA) and volume (V) plays a key role in the energetics of thermoregulation and locomotion, but detailed morphometric measurements of marine mammals are often limited. Thus, SA and V are typically estimated using a series of conical frustrums and cylinders (the truncated cones method), which provides a repeatable but abstracted depiction of morphology. In contrast, digital 3D modeling systems can offer more detailed representations of animal size and shape. We compared the results of the truncated cones method and a 3D modeling system (produced using the open access software Blender) in quantifying the SA and V of both long-finned pilot whales and short-finned pilot whales (Globicephala spp.). We developed a 3D model of pilot whales using measurements and images collected by stranding networks. The 3D model provided a more realistic depiction of pilot whale morphology than the truncated cones method, particularly in the tail stock region where the truncated cones method greatly overestimated both SA and V. The difference between SA and V estimates of the two methods was greater for larger individuals within species, suggesting that as animals become larger, the truncated cones method increasingly overestimated SA and V. Further, the 3D model was more robust to changes in the number of morphometric girth measurements used when estimating SA and V compared to the truncated cones method. Results of this study demonstrate that 3D models can provide realistic depictions of cetacean morphology and can be used to provide more accurate estimates of morphological metrics than geometric models of morphology. The 3D modeling techniques employed in this study could be used in a variety of other applications which require accurate estimates of morphological metrics.