Lewa Wildlife Conservancy, KenyaA tiny bungalow in the central Kenyan bush is the joint operation center for a new kind of weapon in the war against wildlife poaching: information—lots of it.
It’s searingly hot in Kenya’s Lewa Wildlife Conservancy, and the ceiling fan ticks above a wall-mounted touch screen displaying a map with icons of elephants and rhinos scattered throughout the area.
Selvam Velmurugan presses a finger on one of the elephant icons, and an information window pops open above it. “These are live animals roaming the reserve as we speak,” he says. If one of the animals wanders past a border and into a human settlement, he explains, Lewa's managers are sent an alert on their phone so they can react immediately. “It’s like we have a virtual data fence surrounding the reserve. We’ll know exactly where and when the animal has crossed by taking one look at the screen.”
Velmurugan is the manager of philanthropic technology, emerging technology, and machine learning with Vulcan, the Seattle-based company established by Microsoft co-founder Paul Allen. Velmurugan is giving me a demonstration of software created by his team called the Domain Awareness System (DAS), which could revolutionize the way large conservation areas are managed in Africa and around the world.
DAS, he says, is designed to aggregate huge amounts of data in real time—“big data,” as such troves of disparate information are popularly known. In Lewa, the system brings together in a single interactive viewing map GPS readings of animal movements, radio and vehicle trackers to follow anti-poaching teams in real time, camera trap photos, surrounding human settlements where poachers are likely to originate, weather conditions, and more. In this way it gives managers an integrated view of pretty much everything they need to know, minute-by-minute, in what may be a sprawling protected area.
Velmurugan joined Vulcan last year, after stints with Amazon, Netflix, and Groupon because, he says, he wanted “the opportunity to work on the exciting intersection between philanthropy and technology.”
The new software proved its worth in a tryout in Lewa this February when two armed intruders entered a village in the conservancy under cover of darkness and fatally shot two men.
Lewa’s anti-poaching unit, which responds to all security incidents, immediately teamed up with local police and followed the tracks of the men into a village. The conservancy's helicopter was deployed, and a shootout ensued between the authorities and the suspects. In the darkness the suspects managed to get away, but their weapons were seized, and this evidence is assisting in bringing the criminals to book.
“When the attack happened, we knew exactly what was happening on the ground so we could make decisions quickly,” says John Tanui, Lewa’s communications officer, who led the operation from the DAS command center in the bungalow. “The interactive map allowed us to direct our officers to the incident, and we were able to track their every move through their GPS-enabled radios. We had a great overview of the terrain, so we could direct the teams through the bush at night and manage the data from the incident as and when the updates came in. We could see it all visually from the screen. It’s a very good tool.”
Real-time orchestration of responses to such incidents is one of DAS’s strengths. But, Velmurugan says, the tool also has the potential to anticipate poaching incidents before they happen. His team is now feeding data into the system from past poaching incidents in Lewa, such as time of day, day of week, season, vegetation at the time, and rainfall. By crunching all this information, DAS will come up with potential sites where poaching is more likely to take place on any given day. This, Velmurugan assures, will save the reserve significant amounts of time and money by focusing anti-poaching patrols more closely on known poaching hot spots.
“We’re still only scratching the surface of DAS’s capacity,” says Mike Watson, general manager of Lewa, which is part of a 16,000-square-mile conservation area called the Northern Rangelands Trust. Watson says Lewa was chosen to prove out the system because it’s well managed and has embraced innovative approaches to conservation and development.
“It’s hugely valuable to have all our tracked assets—animals, vehicles, and anti-poaching teams—on one screen, which before we couldn’t do,” Watson says. “From a reaction perspective, it makes us much more effective. It gives the commander of the incident a much greater three-dimensional sense of what’s going on on the ground, what assets he has, where they are, where he should be looking to deploy them in reaction to that incident.”
Last November DAS was also deployed at Akagera National Park in Rwanda, and there are plans to introduce it in four more parks by the end of April: Liwonde National Park in Malawi; North Luangwa National Park in Zambia; Wildlife Conservation Society-managed Nouabalé-Ndoki National Park in Republic of Congo; and Singita Grumeti Reserve in Tanzania.
“In total the system will oversee some 40,000 square miles of protected area,” says Ted Schmitt, a Vulcan program manager. Altogether 15 parks, including all those managed by African Parks Network, an NGO that oversees some of the continent's largest and most remote parks, will have it within the next year, “adding tens of thousands more animals under the protection of DAS,” according to Schmitt. Vulcan is donating all the software and assisting with implementation, but the parks themselves will need to build their operation centers and cover the cost of hosting the data.
Meanwhile, Vulcan has teamed up with Save the Elephants, a Nairobi-based conservation organization with a 50-year history of tracking animals in Africa, to release a continent-wide smartphone app based on DAS technology. It allows conservationists, rangers, and reserve managers to keep constant track of some 190 elephants, 46 forest elephants, 45 scimitar-horned oryxes, and 19 endangered Grevy’s zebras and monitor their transboundary movements.
“Accurate data has a critical role to play in conservation,” said Allen, Vulcan’s chairman, in an email. “Rangers deserve more than just dedication and good luck. They need to know real-time what is happening in their parks and where animals are.”
Properly presented and analyzed data can show patterns of wildlife movement and resource use and help enforcement officials do their job more efficiently. The Domain Awareness System and the Save the Elephants app, Allen said, will “ultimately enable rangers who manage thousands of square hectares of terrain to identify and intercept poachers before it’s too late.”
OCEANS OF DATA
Vulcan is now also investigating ways to bring data to bear in combating illegal fishing—one of the biggest conservation problems facing the world’s oceans. A recent study estimated that some 30 percent of wild-caught fish imported to the U.S. were caught illegally, and illegally taken fish are estimated to account for more than 15 percent of the total global catch each year—as much as 26 million tons.
Using a system similar to DAS, Velmurugan's team recently developed a new application to show the location and tracks, in real time, of almost every registered commercial fishing vessel on the ocean. They did this by aggregating the automatic identification signal most ships are required to emit in order to prevent collisions.
“This was the first step to understanding where illegal fishing is happening,” Velmurugan says. “Based on the tracks of a fishing boat, you can actually identify, with expert advice, whether the boat is illegal or not given its geographic location. For instance, fishing trawlers tend to track back and forth in transects. If we’re seeing this behavior within a marine protected area, we’ll know immediately that the boat is fishing illegally.”
Another company, SkyTruth, based in West Virginia, is helping authorities crack down on illegal fishing using identification signals and satellite data. The system started out as a basic tool to monitor illegal fishing boats in the Pacific, then was expanded to enable anyone involved in ocean law enforcement—governments, companies, and advocacy groups—to access the data via the website Global Fishing Watch.
In the past, illegal fishing vessels were able to elude detection by turning off their automatic identification signals, becoming “dark targets.” But increased monitoring by satellites is making it harder to hide.
“The satellite imagery service is one of the fastest growing technology industries in the world,” Velmurugan says. Companies are blanketing the Earth in the low and medium orbits with small-scale “nano” satellites that have high-resolution cameras. These, he says, can “provide an almost real-time ability to say, ‘I want a picture of this area,’ and within a few minutes I can get a picture.”
This January India launched a record 104 satellites, most of which were nano devices belonging to Planet Labs, a San Francisco-based company that will sell the data and images they gather to organizations and agencies looking to monitor the Earth from above.
Sifting through the volumes of raw data from the satellites and using that information to spot illegal fishing boats—long-liners, trawlers, purse seiners, and many other types—is a massive task. So the Vulcan team has devised a way to apply machine-learning algorithms to their illegal fishing detection system, enabling identification of fishing vessels from a satellite image alone. This has increased the system’s detection ability to 93-percent accuracy.
The combination of vast amounts of computing power and access to large amounts of data, Velmurugan says, “is moving us towards a place where we could actually have the machines building their own algorithms on the fly.”
“The technology is going to continue to innovate, and the amount of data that’s going to be generated in the future will be astronomical,” says Vulcan’s Schmitt, adding that it’s now more important than ever to build big data platforms like DAS to host and analyze the floods of information.
Real-time satellite images from space, automatic drones patrolling reserves, cheap sensors on animals such as heart rate monitors to tell if they’re stressed, small cameras in strategic locations, and many other devices that have yet to be invented—all will be used to secure protected areas in the years to come.