Poaching of protected species: the tech battle we must not lose
Image credit: Getty Images
Of all the challenges facing our most glamorous beasts the most emotive is poaching – it is a premeditated act to go and kill an animal. However, gamekeepers are fighting back and are using increasingly sophisticated technology to do so.
By the time you hear a shot, it’s already too late. So says a warden on a large private park in South Africa, where rangers have been waging twilight battles with poachers. In 2014 and preceding years, says David Powrie, too many of the park’s rhinos were slaughtered by criminals who were after their horns. “What do you do?” he shrugs.
Yet this reserve, in the corner of Kruger National Park, is one of the success stories. Since new technology was introduced in 2015, rhino poaching in the private reserve has fallen by 96 per cent in three years.
Old tech methods to deter poachers didn’t curb the trade. South African parks have previously experimented with poisoning or dyeing horns, which are prized in some parts of the world as trophies or as supposed aphrodisiacs or medicines. But criminal gangs don’t care about their end market, and dyes fade, say conservationists. Other methods such as dehorning rhino seems only to boost the value of their horns, and also disturbs the animals.
In this 60,000 hectare park, when poachers breach the perimeter fence rangers have, at best, a small window of opportunity to find them. Poachers usually enter on a full moon or even in full darkness, shoot the rhino, hack off its horns and leave under cover of the night. “That gives us maximum half an hour,” says warden Brandon Jones.
In desperation, the park turned to new technology, which, rather than tracking the rhinos, aims to track the poachers instead. To date it’s been astonishingly successful, albeit on a small scale. “We don’t touch the animals,” says Dimension Data group executive Bruce Watson, a passionate wildlife supporter who’s not a fan of tagging large beasts. “Our solution allows them to roam freely in a safe haven. We monitor the movement of people who are the real threat to this magnificent species.”
This pilot project, ‘Connected Conservation’ – one of many new initiatives worldwide trying to beat poachers, is a collaboration between technology company Dimension Data and Cisco. After consulting with park rangers, the project first installed CCTV and a secure local area network with Wi-Fi hot spots at key points so park staff could communicate and share live video footage via smartphones.This has created a high security ‘net’ around the whole reserve – and its creators believe this could be scaled up to encompass other larger reserves.
This year the second phase of the project has seen the technology expanded to allow security processes to be digitised. Sensors are attached to vehicles entering the reserve so staff are able to track them within the park, until they leave. Wireless technology is being extended within the reserve itself, and perimeter fencing is equipped with acoustic fibre and thermal cameras – protected with tough casing – to feed back to the park’s headquarters and alert rangers if there’s a break-in. Magnetic sensors have been buried underground, and can detect when weapons cross the boundary.
Project managers believe this technology can eventually be expanded to other continents. “I believe our job is to create a safe place for the rhino and other animals to roam freely,” says conservationist Dave Varty, who’s involved in the project.
GPS-trackable imitation eggs
Is technology the answer to the illegal wildlife trade? Yes and no, say conservationists. “It really isn’t a panacea,” says Dr Paul Glover-Kapfer, conservation technology adviser at WWF-UK. It’s one thing to monitor wildlife – it’s another to be able to act on the vast amount of information. “What we need, rather than shiny new tech, is proven, off-the-shelf solutions, basic equipment and also boots on the ground. One of our major challenges is that we don’t have sufficient funding to have and train enough rangers.”
Even when rangers have the necessary intelligence – delivered by technology – they’re not always equipped or trained to deal with criminal gangs. When mini drones armed with surveillance cameras were flown over national parks in Zimbabwe and Malawi, poaching decreased, largely due to a deterrent effect. But this pilot didn’t result in any arrests of poachers by rangers. This isn’t their fault, say conservationists.
“Rangers are too frequently facing dangerous and challenging situations and they all too often don’t have the right equipment and training to be able to stop these ruthless criminals,” says Heather Sohl, chief adviser for wildlife at WWF. “In order to halt the poaching we need governments across the world to ensure there are enough rangers with the necessary skills, equipment and insurance on the front line of defence.”
Operating in the vast acreages of national parks and remote terrain presents an enormous technological challenge, says Dr Glover-Kapfer, because the bush and technology don’t go hand in hand. “These are the challenges of working outside: it’s too hot, too cold, too wet, too dry. Vandalism is common. We are up against intelligent organised criminal syndicates; they come up with their own technological solutions to our technology.”
While poachers might be tech-savvy, most people working in the field are biologists and conservationists who don’t have technology backgrounds, says David Northmore, vice president at database technology provider MarkLogic, which is working on anti-poaching technology. “Modern poachers are incredibly well resourced; using helicopters, tracking equipment and sophisticated weapons to poach elephants and rhinos among other species.”
Sometimes criminal gangs have hacked signals from animals’ tagging devices, leading them directly to the wildlife, or they’ve listened to rangers’ communications over VHF radio. Satellite tracking has nonetheless provided conservationists with valuable information on animals’ whereabouts and habits, but often tagging collars are bulky and intrusive, battery life and range are limited, or they’re too fragile.
And although much of the technology used in the field has been in circulation for years, it’s still early days in terms of harnessing its capabilities, says Stephanie O’Donnell, community manager at WildLabs.net, an online community which shares ideas, tools and technology to fight conservation challenges. “Technology does have a huge role to play,” she says. “There are so many opportunities and the solutions are evolving. Look at camera trapping – a tool that’s been around for decades, but is still one of our most innovative pieces of technology. And tags are becoming smaller, cheaper and more efficient. Satellites can provide incredible data. But we need the systems – the machine learning, the artificial intelligence, to be able to make sense of it all.”
Northmore agrees, saying this is a gap that must be plugged.
There are a huge range of initiatives under way and some off-the-shelf products already in use, says Dr Glover-Kapfer. “We have a toolbox of available technology, which we can then apply to unique contexts. But the most amazing thing about this work is how complex it is. What works in one place won’t necessarily work elsewhere.”
So what new tools are evolving? Great strides are being made in sensor technology. As well as locating the animals, acoustic sensors are becoming cheaper and more robust, allowing conservationists to eavesdrop remotely. “You can listen in to animal calls, to gunshots, to use of chainsaws,” says Dr Glover-Kapfner. “But you will need the help of AI to decipher the data – imagine how much of it there will be.”
Sensors need to be small, cheap to make and they need a long battery life – or ideally no battery at all – says Marion Moreau, head of Sigfox Foundation. For the last year and a half, the foundation has been tracking rhinos in Zimbabwe using Sigfox’s low-power, long-range network and sensors in a pilot project. “We chose Sigfox technology because it’s less intrusive, cheaper (sensors cost $50), batteries last up to five years and it’s more secure because protocols are encrypted. They are very robust and secure – you can’t hack them.”
Vets working with the project ‘Now Rhinos Speak’ insert a small tracker – about the size of a human finger – into a rhino’s horn while the animal is sedated. Each sensor emits three signals a day using GPS location. Geofencing works alongside location tracking by creating a virtual geographic boundary “so if the rhinos leave their usual stomping ground and approach a dangerous or populated area, the sensor sends a warning to a ranger’s smartphone that it might be at risk”, says Moreau.
Now Rhinos Speak has secured money to produce 20 more trackers, and nations such as Kenya, Tanzania, Botswana and South Africa have requested the technology. By the end of the year, Moreau hopes to be trialling solar-powered sensors. “We want as little human intervention as possible,” she says. “This year we want to reach our goal of producing very affordable open source solutions for the whole world.”
New prototypes will experiment with the addition of an accelerometer, which monitors how the animals move, as opposed to where they are. “Once enough data has been collected, it will help generate an algorithm which, together with GPS data, can better predict the risks rhinos face and improve our understanding of how they respond to threats,” she explains.
In a separate project, Dutch social enterprise Sensing Clues, working with MarkLogic, has developed an early warning system for rangers protecting elephants from poachers in Kenya. A data hub framework encompasses a number of different technologies – smart IoT sensors to detect human activity in remote places, storage of complex data and analytical intelligence for spatial awareness and response coordination technologies. Sound sensors can distinguish between different types of vehicles and weapons and light sensors can pick up humans using artificial lights in vast terrains. “With this 360-degree view of data they can now easily see when people are entering restricted areas,” says Northmore. This technology is also being used in Sweden, where poaching of wolves is a serious problem.
Effective technology, says Dr Glover-Kapfer, should be easy to use. “Lots of technology requires a fair amount of skill – you might need engineers on staff to deploy, troubleshoot and maintain. But democratisation of technology is getting better and we’re getting closer to the point where we could have a ‘plug and play’ toolbox that anybody can pick up and use.”
Anglers crowdsource Fish Population data
A range of companies including Facebook, Google, Microsoft and Cisco are throwing their hats into the ring and collaborating with conservationists, specialists and entrepreneurs – many of which swap ideas on WildLabs.net.
“Camera traps (motion-triggered cameras) have really transformed wildlife monitoring over the last 15 years,” says Dr Glover-Kapfer. “They are ubiquitous, indispensable and they provide this amazing, engaging content.”
However, the challenge of sifting through thousands of images can be overwhelming, as WildLabs.net’s O’Donnell points out. An ongoing project to monitor lions in the Serengeti National Park has created millions of camera-trap images. Identifying species is a complicated process, even for humans.
It’s here that machine learning and visual recognition technology could lend a hand to process mountains of visual wildlife data. With this in mind, the conservation technology unit of the Zoological Society of London (ZSL) is working with Google Cloud in San Francisco and San Diego Zoo, using the zoos’ vast archives of data. They are developing platforms for training machine learning algorithms which will be able to identify individual species from thousands of images, test recognition models and refine machine learning algorithms, which could then be applied to fieldwork. One of ZSL’s aims is to make artificial intelligence available to smaller groups and conservationists who lack the expertise to train models for their own particular needs.
You can now even spot animals from space, “which is stunning,” says Dr Glover-Kapfer. Polar bears, whales, penguins, seals, wildebeest and more have all been captured by satellite remote sensing. Satellites are now cheaper to deploy and offer increasingly higher-resolution images. This could allow conservationists to count and monitor wildlife without disturbing it. “And this is only going to improve,” he adds, although researchers will face similar problems with coping with quantities of data.
Drones also have a role to play in tackling poaching and tracking rare breeds, although they are only as intelligent as the sensors they carry. “They will eventually become an incredible, powerful tool,” says Dr Glover-Kapfer. “It’s just a matter of time.” Equipped with thermal imaging cameras, drones are able to monitor orangutans, for instance, with greater precision than human trackers can. “Rather than counting nests, they can register the heat signature of the bodies, so you can count the animals,” he says.
In fact one of the most effective is a spatial monitoring and reporting tool (aka Smart), developed by nine leading conservation organisations including ZSL and WWF. “Smart has exploded onto the scene,” says Dr Glover-Kapfner, “It’s incredibly effective.” Created for rangers and park managers in 2012, it’s an adaptive system which helps train staff, gather information on wildlife and threats, and analyse data to come up with better protection plans. Open source and freely available, the law enforcement management system has been deployed on the frontlines in more than 600 sites across protected areas in 55 countries. In its first year in one wildlife reserve in Nepal, it led to 30 arrests and the removal of 22 hunting camps. Recent upgrades allow for faster communications and analysis.
If ever effective tools were required against illegal wildlife trade, it’s now. The rate of trafficking has spiked over previous years as people’s appetites for exotic pets, bush meat, jewellery and traditional medicines grow. “Time is against us,” say the conservationists.
With so much innovation, do we now have the tools to prevent and crush illegal trade in wildlife? And can these tools be used to monitor and save endangered species?
“Ultimately the unifying theme with all these various technologies – primarily hardware sensors that allow us to collect data – is that we must be able to use this information, with the help of artificial intelligence or machine learning, to predict what will happen,” says Dr Glover-Kapfer.
Technology must be affordable, reliable and easy to use. It’s all about integrating data from different sources, says Northmore “to gain a 360-degree view of the situation”.
There is no shortage of initiatives – from meet-up workshops hosted by ZSL, to incubator programmes and support for community platforms such as WildLabs.net, to field initiatives such as Data Dimension’s Connected Conservation.
Every year, leading conservation groups, including London Zoo, welcome coders and programmers for a short session – Zoohackathons (zoohackathon.com) – a US initiative aimed at developing solutions to problems of trafficking. Ultimately the conservation community needs more professionals with the ability to help rangers and conservationists do their jobs.
“An engineering skillset can make some of the most valuable contributions to conservation,” says Dr Glover-Kapfer. “It would be lovely to get more engineers to sign up to WildLabs community and participate alongside conservationists in developing solutions to problems. We don’t have nearly enough.”