An Analysis of Utilizing the Leatherback’s Pineal Spot for Photo-identification

Abstract

Leatherbacks are one of the most endangered species of sea turtles. Their global population size had decreased dramatically over the past several decades. Difficulties in applying long-term marking methods on leatherbacks, have significantly hindered our ability to generate an accurate population estimate and to track population changes on the scale of individuals. External marking methods are plagued by high rates of tag loss which compromise the utility of tagging as a means of long-term identification. While having substantially better retention rates, the high cost PIT tags and the specialized equipment necessary to detect them, have limited the expansion of this technique. This project outlined the need for innovative approaches to identifying turtles and sought to resolve some of the issues plaguing traditional tagging protocols by reinvestigating the feasibility of using photo-identification to recognize individual leatherbacks. Pioneering studies of pink spot photo-identification by McDonald and Dutton (1996) showed that the pineal spot of the leatherback is distinct enough to be used as a unique identifier. This research investigated whether individual turtles could be recognized by their “pink spot” when a large sample (~400) was drawn from a very large nesting population (~ 3,000 nesting annually). It also sought to determine whether identification could be automated using photo-identification software. The results indicated that the pineal “pink” spot’s form has sufficient variation to be used as a unique identifier in photo-identification studies. Using the Scale Invariant Feature Transform (SIFT) function we were able to successfully automate identification. Through a cascade filtering approach the program was able to achieve one hundred percent matching accuracy, while eliminating the possibility of false negatives. Further studies are required to examine the degree of deviation that may occur within the pink spot over time, and to refine our methodologies to expand the use of this method amongst leatherback researchers.