Visitors to the stretch of coastline from Donegal to Antrim, Northern Ireland, are often treated to the sight of bottlenose dolphins leaping from the sea surface. Pictures of their acrobatics accrue thousands of likes on social media, but amateur photographers are often unaware that their images are generating powerful data.
Last year, a study from the University of Belfast used social media to get a better view of the community and behaviour of Northern Ireland’s bottlenose dolphin population. From 2006 onwards, the team collected images, along with the location and date information, from Facebook, Twitter and Flickr. The distinct markings and fins were used to identify individuals to create a marine ‘social network’ for Northern Ireland’s dolphin community. When citizens re-sighted known individuals, the team could visualise the dolphin’s movements, behaviour and community structure. ‘Our study gave evidence for year-round occurrence of bottlenose dolphins, and that certain individuals were coming back year after year. That indicated how important the area was for the this community,’ says Suzanne Beck, one of the study’s researchers. Their work showed that Northern Ireland’s bottlenose dolphins are part of a larger coastal population that range throughout the Northern Irish and Irish waters. The dolphin ‘social network’ is now being continually monitored and used to inform policymakers, helping to guide decisions on development projects that could impact their migratory routes.
In recent years, social media and citizen science has made the impossible possible for many marine researchers. ‘A major challenge that we have is that there are just a few professional researchers spread across a massive ocean,’ says Simon Pierce, director of the Marine Megafauna Foundation (MMF).
‘On top of this you have species that migrate hundreds of kilometres every year to different breeding or feeding areas. It’s just a huge ask for one research group to try and cover an entire species’ or populations’ range,’ adds Beck. ‘It’s often impossible for a standalone project or a small NGO to undertake. That’s where citizen science comes in.’
Hobbyists and divers are becoming nodes for data generation, generating content that can be leveraged for a collective conservation goal. The approach is particularly useful for monitoring endangered species, such as the whale shark – listed in the IUCN Red List of Threatened Species as Vulnerable to Extinction. The Wildbook for Whale Sharks library (supported by MMF) is a visual database of whale shark encounters; divers upload images of encountered sharks, which are then put through pattern recognition software that identifies individuals based on their unique stripes and spots. So far, more than 13,000 individual whale sharks have been identified from around 80,000 reported sightings. The database, or ‘social network’ of whale sharks, is being used by local, regional and international conservation authorities to map the world’s population, to visualise their range and migratory routes and to guide protection efforts. Pierce has just embarked on a new study in Thailand using this citizen science data – it is the first data ever generated on Thailand’s whale shark population, and is already gleaning information on their migration patterns.
A parallel project is underway for manta ray conservation. Each ray has a unique pattern of spots on its belly. Social media users can upload pictures where the markings are visible, which are picked up by MantaMatcher through hashtags. Around 12,000 individuals have been identified to date.
The Turtle ID project in New Caledonia, Sri Lanka, is creating a database of locally spotted sea turtles by pulling in images discovered through hashtag searches. The population size, species distribution and turtles’ feeding behaviour is being monitored through the project. Turtle ID, MantaMatcher and Wildbook for Whale Sharks are all possible due to growingly sophisticated machine learning software: Turtle ID, for example, uses open-source software called I3S Pattern, which identifies individuals based on machine learning principles. When a usable image is discovered, reference points are taken from the turtles’ nose tip, the inner edge of the eye and the furthest scale. The software then outlines other identification marks, before showing which turtles have the closest match from its database – or ‘social network’ – of New Caledonian turtles. Artificial intelligence tools are adding new tools to the armoury for Wildbook for Whale Sharks: ‘There’s now an ‘intelligent agent’ searching YouTube for encounters: each time a video of a whale shark is uploaded, the AI tool will add the shark to the database to try to identify the individual,’ says Pierce.
Yet, some researchers urge caution. Quests for the perfect photograph might disrupt the natural behaviour of marine animals: studies show that as much as 80 per cent of humpback whales deviate from their trajectory when boats approach; sea turtles spend markedly less time breathing and feeding on reefs when divers approach; and, when tourists are not in-the-know, they might violate protection measures. In New Caledonia, for example, it is illegal to approach sea turtles at distances below ten meters. Pierce believes that for now, citizen science projects have been entirely positive: ‘There’s a hypothetical chance of disturbing animals, but the contributors all love these animals – it’s easy to educate and train them to take photos without causing any stress.’
A more difficult challenge is to overcome the possibility of skewed or incomplete data. Tourists and divers often venture out on fair-weather days, and inevitably, not every social media user is capable of taking usable pictures. Distributing guidelines to those who want to participate in citizen science studies could minimise natural behaviour disruption and improve the quality of images being generated.
Citizen science, in an era where most are generating content for social media platforms, is increasingly seen as a bridge between science and society. Beck believes contributing to science can be a trigger for meaningful engagement with nature and protection efforts: ‘It gives people a huge sense of pride with their own patch of coastline, and when they contribute it creates a sense of purpose and environmental stewardship.’