Video: Critter cams reveal animal secrets
Imagine if you could get inside the mind of an animal, share its perceptions and see exactly how it responds to other animals and its environment. It’s a tall order, but now we can do the next best thing: accompany animals unnoticed as they go about their lives.
Video cameras can now be made small enough to be mounted on animals unobtrusively, while even tiny memory cards have enough capacity to record hours of footage as the animals roam. In the past, animal-borne video and environmental data collection systems – better known as – could only be fitted to large marine mammals whose mass, coupled with the buoyancy of the water, allows them to lug around heavy, bulky camera units without serious impediment.
Advertisement
More recently, terrestrial animals have been equipped with video cameras, but except for large ones such as lions and grizzly bears, the limit on the size of the batteries and memory devices that they can carry has made the footage too brief to be helpful for scientists, and its use has largely been restricted to . The cameras were also bulky enough to affect the animals’ behaviour, reducing the usefulness of the footage even further.
Now that is changing. Advances in cameras, recording units and, especially, batteries mean video units can now be attached to much smaller animals, including deer and even birds, without changing their behaviour but with enough memory and power to chug away for long periods of time. “We can study so many more species,” says Joshua Millspaugh, a wildlife ecologist at the University of Missouri in Columbia. “We can almost walk alongside the animal and see what it’s seeing and hear what it’s hearing. It’s starting to get really exciting.”
“We can almost walk alongside the animal and see what it’s seeing, hear what it hears”
The plummeting cost of cameras and memory also means that many more animals can be equipped, allowing enough data to be collected for statistical analysis. “The attitude toward this technology has changed,” says Christian Rutz, a behavioural ecologist at the University of Oxford, who predicts that costs could fall much further, from thousands of pounds today to below £200 in the near future. “Many people now perceive it as a valid tool for doing quantitative, hypothesis-driven research, rather than one-offs that might produce a snippet of footage for National Geographic.”
Critter cams are already yielding results. Researchers studying the spread of chronic wasting disease – the deer version of mad cow disease – need to know how frequently deer interact in the wild. This is nearly impossible to quantify by trailing deer through the forest, or sitting in a hideout hoping to catch a glimpse of an interaction. So Millspaugh and his colleagues have fitted deer with tiny cameras on their antlers, wired to a body-mounted battery pack. Their shows that that deer interact with each other – and thus have a chance to pass on the disease – more often than previously thought.
The technology to do such studies is getting smaller all the time, allowing increasingly sophisticated analyses. In 2006, Rutz and his colleagues mounted 14-gram video backpacks onto the tail feathers of New Caledonian crows. These birds are famous for their abilities to use tools, and Rutz’s team wanted to see how they use tools in the wild (Science, ). To keep the camera unit as light as possible – crucial if the crows are to behave naturally – Rutz’s team did not include a way to store the video data on board the device. Instead, they built in a radio transmitter, which is much lighter than a memory system, to broadcast the video stream to a nearby antenna. The difficulty is the transmitters generally need a clear signal path to an antenna no more than a few hundred metres away, limiting the amount of information they can send.
To get around this, the team plans to take advantage of new computer processors and solid-state memory cards, which are small enough to make data storage practical. They are working to incorporate these into that will store several hours of video. By attaching a homing beacon to the crows, the researchers can then track them and retrieve the footage. Currently these weigh 17 grams, but they need to get them down to 14 grams to be sure that the device won’t hinder the crows. “Shaving off these extra few grams is quite difficult,” says Rutz.
He hopes that the ability to gather much more footage will allow them to measure the extent to which the tools help the birds to forage and to see whether individual crows prefer different kinds of tools. It is known that the crows use a range of different tools, but not whether individuals have preferences.
It’s not just about the size of the equipment, though: Rutz and others are also using some tricks to maximise the amount of useful information they can record. For example, the new crow video packs will incorporate a delay circuit so that the video doesn’t start rolling for 24 to 48 hours. That way, says Rutz’s collaborator Lucas Bluff, “the footage we get is of a crow doing what a crow does, not a bird that has just been captured”. Similarly, Millspaugh’s deer cams carry motion sensors that turn the camera off when the animal has not moved for a while, sparing him hours of worthless video from a sleeping deer.
An animal’s-eye view can also be combined with other data – such as GPS location trackers – to identify what stimuli might have led to certain behaviours, such as a change in direction. “You get the video in a spatial context,” says Rutz. “You see what an animal did where and when, and whether it did it in the presence of others. There is huge research potential by marrying these two technologies. We think that is the future of wildlife research.”
Tim Guilford, an animal behaviourist also at the University of Oxford, may help take us into that future. For several years, he has been fitting trained homing pigeons with GPS recorders. By correlating the pigeons’ changes in course with landscape features, he has gained insights into which features guide the birds as they navigate. Now, with colleague Matthew Collett, he plans to combine this GPS record with a head-mounted video camera so they can see what the pigeon sees as it flies along and makes decisions. Pigeons move their heads a lot when flying, so video from a head-mounted camera provides a good guess at what they are looking at.
Glimpses of animals in action also allow new insights into how their bodies work. Earlier this year Graham Taylor and his colleagues at the University of Oxford mounted a video camera aboard a trained steppe eagle to record close-up images of how its wing feathers moved during flight to change the aerodynamics. During landing, they found, the feathers on the underside of the wing act much like the flap on the leading edge of an aircraft wing, increasing lift so that the bird can slow down (The Journal of Experimental Biology, ).
Hopefully, the scientific pay-offs of animal reality TV will be accompanied by another, less tangible benefit. “People seem to gravitate to images from the perspective of an animal as it goes about its life,” says Greg Marshall, a conservation biologist with the National Geographic Society in Washington DC. “They seem to understand the challenges the animals face in a more direct and visceral way.” If that translates into a stronger commitment to conservation, critter cams could turn out to be a very good thing.
Making sense of it all
As memory storage and battery power increase, the data stream gathered by animal cameras will swell from a trickle to a torrent. Researchers will have to find some way to make sense of it all.
To do this at present, a researcher must painstakingly analyse each frame individually. That will become impossible as the quantity of information rises. “We joke that there aren’t enough graduate students on the planet,” says Greg Marshall, a conservation biologist with the National Geographic Society in Washington DC. Some form of automated analysis is on everyone’s wish list.
The Monterey Bay Aquarium Research Institute (MBARI) in California monitors marine life at the bottom of the ocean using video cameras mounted on remote control vehicles. It has developed an automated that uses colour, shape and persistence through several frames to flag moments when “something interesting” appears. The system can distinguish likely looking shapes from random flotsam, but it still needs a person to verify that the object is an animal and identify it, says Judith Connor of MBARI.
Many technologists are working on the similar problem of automatically scanning footage from security cameras. But sifting through sequences recorded by a stationary camera is easier than making sense of the jumble of plants and animals in an ever-changing forest scene. Even DARPA – the US Defense Department’s research wing – is finding it difficult. It presented its efforts into automated video analysis at a organised by the National Geographic Society in October. “We all left feeling that it is going to be our biggest challenge,” says Marshall.
