The Long and the Short of It: A Simple Photograph-based Approach for Discriminating between Free-ranging Long-finned and Short-finned Pilot Whales off the East Coast of the United States
Figure 2. Map depicting the study area, range overlap, and genetic samples collected during the years 2004-07.
Figure 3. Reference lines (solid lines) used to define the six morphometric features and relative measurements collected (dotted lines) on an image of a free-ranging pilot whale used for species discrimination.
Figure 4. The four categories analyzed for pigmentation. The pilot whale in this example displays the presence of pigmentation in all four areas: melon patch, post-orbital eye blaze, cape, and saddle patch.
Long-finned (Globicephala melas) and short-finned (G. macrorhynchus) pilot whales look similar at sea. Described pigmentation characters include a post-orbital eye blaze, a post-dorsal saddle patch, and a cape that connects the eye stripe and the saddle. Although these two species are readily distinguishable using osteological characters like tooth counts and skull morphology or by genetic analyses, they are generally considered not to be predictably distinguishable at sea by physical characteristics. In some areas (e.g., the western North Atlantic), their ranges overlap, thus, this difficulty in identifying individuals to species limits at-sea studies.
In the western North Atlantic, long-finned pilot whales have an anti-tropical distribution with a southern limit of North Carolina, and short-finned pilot whales are found in tropical waters with a northern limit of New Jersey. The shelf edge/slope water region is an important habitat for both species of pilot whales and, in the western North Atlantic, their ranges overlap during summer from approximately 35°N (Cape Hatteras, North Carolina) to 39°N (east of Delaware Bay) (Fig. 2).
Understanding potential anthropogenic impacts (e.g., fisheries) to both species of pilot whales in the western North Atlantic as reviewed in the NMFS stock assessment reports (SARs) is important to properly manage these populations. Their similarity of appearance, significant range overlap, and the inability to distinguish between them during assessment cruises add uncertainty when describing the geographic range, estimating the minimum population size and human-induced mortality and serious injury rates, and calculating the potential biological removal (PBR) level, which is required for each stock. To date, all SARs of western North Atlantic pilot whales are inadequate; species-specific abundance estimates have not been produced due to the uncertainty of species identity within a large proportion of their putative summertime ranges.
In this study, we developed a photograph-based approach to distinguish these two species of pilot whales in the western North Atlantic. We collected and examined photographic images of pilot whales coincident with genetic sampling during five summer research cruises conducted in U.S. waters of the western North Atlantic and from one stranded animal during the years 2004-07. We examined the utility of physical features that are detectable in photographs such as relative shape or coloration as useful discriminating characters.
For morphometric analyses, 11 combinations of ratios calculated from six morphological features along with a three-level size-class covariate (small, medium, large) accounting for sexual dimorphism and allometric growth rate were used to test discrimination functions for species identification (Fig. 3). Two datasets were analyzed: the training set (non-genetically sampled animals) used to calculate parameter estimation and the validation set (genetically sampled animals; see Fig. 2) used to assess the model’s predictive power. The resulting discrimination function equation (± SE) used to predict species identification is:
Pigmentation was analyzed as presence/absence of four pigmentation characters: melon patch, post-orbital eye blaze, cape, and saddle patch (Fig. 4). All four characters were present in 100% of the short-finned pilot whales, explaining an overall lighter appearance compared to long-finned pilot whales in which melon patch, blaze, and saddle patch were present 6%, 68%, and 50%, respectively, and the cape was completely absent (Figs. 5a, b).
Figure 5a (left). Classic pigmentation characteristics on two short-finned pilot whales: melon patch, blaze, cape, and saddle patch extending from the cape down the dorsal caudal peduncle;
5b (right). A long-finned pilot whale displaying both the absence of the cape and a saddle patch with a distinct boundary located just posterior to the dorsal fin.
As a result of this study, two procedures were designed to identify pilot whales at sea in the western North Atlantic under varying field conditions. If a sighting event occurs during sufficient daylight and the lighting is such that pigmentation contrast is easily discernable (i.e., sunny conditions and minimal glare) and capturing acceptable images for morphometric discrimination is not possible, analysis of the pigmentation is likely sufficient for determining species. However, during times with difficult lighting conditions (i.e., glare, sunrise, sunset, or overcast), a morphometric analysis should be applied. To achieve the most accurate species identification, the two approaches should be used together to provide multiple pieces of evidence for determining the species identity.
Until now, many cetacean biologists thought that reliable determinations of pilot whale species could not be made at sea in areas where both species were found. In this study, image utility was explored, image protocols for analysis were defined, and results showed that a species designation for free-ranging western North Atlantic pilot whales was quite reliable when implementing a simple, inexpensive method. This method can easily be incorporated into future cetacean abundance surveys off the East Coast, which could then result in separate abundance estimates of short-finned and long-finned pilot whales.
