What is The Leatherback Project Photo Identification Initiative?

Our overarching goal is to improve estimates of leatherback sea turtle population size, internesting movements, migration patterns, fisheries interactions and nesting behavior while simultaneously providing a platform for community outreach, fundraising and conservation campaigns.

How does it work?

We aim to combine photographs, citizen science, and computer algorithms to identify individual leatherbacks wherever they are, and to combine these positive identifications to understand patterns of range-wide connectivity. We rely on contributions from colleagues ‘on-the-ground’ – or in the water – and try to maximize the value of face-to-face interactions with leatherbacks around the world. The dataset collected by various groups throughout the globe of individual leatherbacks will be the first used to develop computer algorithms for automatic individual identification using the white spots on the sides of the face of a leatherback. Identification using pink spots will be combined with identification using white spots for maximum confidence in the results. This is a pilot project and is in its trial phase.

Why photo identification?

Identification of individual animals allows us to understand fundamental aspects of biology, as well as track the status of populations over time. For marine turtles, widely accepted forms of identification include PIT tagging, where a chip with a 9-digit identification code the size of a grain of rice is inserted into the shoulder of each turtle. PIT tag scanners are expensive (up to $1600), causing this method for identifying individuals, although very important, to exclude all non-biologists who may be interested and able to participate in conservation efforts. Metal flipper tags, while less expensive, typically fall off of leatherbacks within a few years, limiting their usefulness as long-term tools for tracking individual turtles over time. For large conservaiton projects, these hurdles are not a problem, but for many local, grassroots organizations, these costs completely prevent them from collecting information on the individual turtles that nest on the beaches they monitor.

In contrast, photo identification of the leatherback sea turtle can be useful in providing information on fine-scale population movements, fisheries interactions, and nesting events on beaches where tagging has not been feasible. A local family on the coast of Nicaragua, for example, may see sporadic leatherback nesting in their backyard, be interested in helping with conservation efforts, and may also have access to a smart phone. A picture of a turtle’s face could be taken, and a data point could be collected very easily, while bypassing the cost of PIT tagging, scanning, and the training required to do so. Photos can be taken without flash and with red lights when possible, to minimally affect the turtles’ nesting process.

There have been several studies focused on identifying individual leatherbacks using photos of the pink spot, but none that have used the white spots on the sides of the turtles’ faces to distinguish between individuals. Developing a program that can analyze the spots on the face instead of or as well as the pink spot is important for several reasons. One, if we are able to acquire pictures of leatherbacks from previous years, it is likely that the side of the face is captured more frequently than the pink spot would be. Getting a good picture of the pink spot is more challenging and specific, and historically these pictures were only taken if someone was specifically requesting them. Developing software that can evaluate facial spots will likely allow us to use the largest number of pictures that will inevitably be submitted or available for this effort.

What we propose will hopefully be more universally useful and will have a higher overlap with photos that tourists or local people will take, even before hearing about this project. We will include images of the pink spot on the top of the turtles’ heads as well as the white spots on the sides in order to make use of whatever images we are able to acquire from previous years and to make individual identification as robust a process as possible. The Leatherback Project is collaborating with Wild Me (https://www.wildme.org) to create the identification algorithms. Presently, we are testing the feasibility of this identification method and working on developing and implementation of the necessary computer algorithms. Our approach will be to use both the pink and white spots on a leatherback turtle’s head to identify individuals. This process through Wild Me is quoted to cost $5,000, but once implemented, the global photo ID database has potential to save thousands of conservation dollars annually typically spent on expensive tagging equipment. Every little bit helps us reach our goals!