Conservation efforts need data. To understand how a species is faring, it’s essential to have information on roughly how many members of that species are living and where they’re living.
But this kind of information is fiendishly tricky and expensive to get. Wilderness areas cover huge tracts of land, large segments of which can be difficult or impossible to access. Methods to estimate population densities rely on expensive equipment and staff time, and conservation is desperately underfunded.
A paper in Current Biology this week proposes a cheap alternative that could work in certain contexts: tourist photos. Researchers found that data gathered from tourist images in Botswana provided comparable results to more traditional survey methods—and at a much lower cost.
Data there for the taking
In tourist-heavy conservation areas, large groups of people are out in the field daily, looking for and recording the presence of animals. Lead author Kasim Rafiq realized how much useful information was being gathered on these trips when safari guides told him that they’d seen a one-eared leopard named Pavarotti just that morning. Rafiq had been trying to find Pavarotti for months.
So Rafiq and his colleagues devised a way to get hold of that data. They approached tourist groups staying at a safari lodge in Botswana and asked them if they would be interested in contributing their photos of large carnivores—lion, leopard, wild dog, cheetah, and spotted hyena—to the project. Out of 26 groups approached, all had members who agreed, suggesting that interest in this kind of work would probably be high.
People within the group were given GPS loggers that would record where the tourist groups had been on their game drives, which could be as much as four or five hours long. The GPS logs would then be matched up with the timestamped photos from the tourists’ cameras, collected at the end of their stay.
The researchers came up with an ingenious way to deal with the problem that no camera timestamp is ever correct—tourists were asked to photograph a computer screen displaying the correct time, which was then used to correct the camera timestamp in the data.
Overlapping estimates
Rafiq and his colleagues identified the individual animals in the tourist photos using physical markings, and they then used the data to extrapolate estimates of population density (how many individuals were estimated to be living within 100km2).
The researchers next compared these results to estimates that they calculated using three more traditional survey methods: camera traps, spoor tracking, and call-ins, which attract animals by playing noises through a speaker.
The different methods all produced different ranges of population density estimates, but importantly, the estimate ranges from the tourist photos overlapped with the ranges from traditional methods, suggesting that the tourist picture results were comparable. For cheetah, the tourist method was the only one to even identify that cheetah were living within the study area—although there weren’t enough individuals to calculate a density estimate.
Obviously, crowdsourced data like this comes at a much lower cost than other methods that rely on researchers collecting data themselves. It was only slightly cheaper than the spoor survey method, but it cost less than half of the call-in method and a quarter of even the cheapest estimate for camera trapping (one estimate that assumed being able to re-use previously purchased camera equipment).
Good for lions, but probably not snakes
One potential concern with the tourist method is that safari guides might have a good idea of the favorite haunts of certain species, and thus be able to reliably find those spots. That would mean that the populations are denser in the place that’s being sampled by the citizen scientists, compared to other spots within the overall study area.
This is a problem because the methods that produce the density estimate assume that any site is representative of the whole study area—so, the place where the animals were found is assumed to be average, not extra-dense. Rafiq and his colleagues made some adjustments to account for this, but the researchers point to this as a potential weakness of this method.
Then there’s the obvious limitation that tourists are generally interested mainly, or exclusively, in charismatic megafauna. The lions get a lot of love; less exciting species, not so much. But all methods have their limitations. Tourists’ cameras are just an additional potential tool available to conservationists—and using them to survey certain species could free up resources to focus efforts elsewhere.
It could also be possible to reduce the research costs of this method even further by automating parts of the data sorting process, Rafiq notes in The Conversation: “If the efforts of tourists were paired with AI that could process millions of images quickly, conservationists could have a simple and low-cost method for monitoring wildlife.”
Current Biology, 2018. DOI: 10.1016/j.cub.2019.05.056 (About DOIs).
https://arstechnica.com/?p=1541725