This week, a major summit began in London aimed at tacking the $19 billion a year trade in illegal wildlife trade, which is causing the populations of some of the most charismatic and endangered species on Earth to crash to critical levels. Last year, 1,004 rhinos were killed in South Africa alone, and more than 25,000 elephants were killed by poachers across Africa. That is three elephants every hour, every day. If there is one thing that every participant at the summit can agree on, it is that the current approach is not working.
Many conservation organisations have decided that the answer to combating poaching is simply to throw technology at the problem – such as satellite tracking collars – and see if it works. The result has been a tremendous waste of money and the introduction of useless technology into Africa. In the meantime, there has been a huge increase in poaching over the past four years.
So what can we do? Over the past two years, our team at the University of Maryland’s Institute for Advanced Computer Studies has created the world’s first comprehensive analytical model of poaching behaviour in Africa.
Using satellite imagery, exquisite mathematics and complex algorithms, we have been able to learn how animals, rangers and poachers move through space and time. For example, we try to find out what the local environment looked like when a poaching incident occurred. We gather as much data as possible. Was it rainy, windy, full moon or no moon? What was the day of the week? Were other criminal acts happening that day, or the day before or after? Where were the rangers that night? Were they sent to the opposite side of the reserve? Did we know where the rhinos were supposed to be?
We have learned that we do not have to find poachers: we just need to find their prey. By sending rangers to where the animals are, we can protect them. It sounds simple, but up to now, it has not been happening. The rangers believe they must patrol the entire reserve, so every day they move to different areas.
Why? Because that is what they are told to do. They assume poachings are random events and thus, are equally likely to occur in any given place in the reserve. This is not so.
Last May, we added unmanned aircraft systems to our armoury. As with any other technology, UAS are only a tool; the key is to know where to fly them.
To test our work, we began flying night flights at Olifants West, in the Balule Reserve in South Africa, just west of Kruger National Park, whose black rhinos are under severe threat.
Our models generate both precise flight paths for the drones and strategic deployment plans for the rangers. The mathematics tells us where the rhinos are likely to be – and the modus operandi of the poachers in this area.
We know to fly the UAS over the rhinos and to watch certain areas for poachers moving into a park. Because the model has told us where to position the rangers, we have a very high likelihood of intercepting the poachers before they can reach the animal.
Of course, using a UAS to detect a poacher 50 miles away will do nothing to save a rhino if you cannot get an armed ranger to respond to that sighting before they are killed. The most important factor is how fast a ranger can move, at night and in heavy bush, to reach a poacher before they get to an animal.
Over the last six months, we have identified some interesting trends. At Balule, which is a small reserve compared to some others in South Africa, rhinos are usually killed within 160 yards of a paved road. This is because the poachers are driving around the perimeter fence late in the afternoon, looking for animals near the fence that can be seen from the road.
As soon as the sun sets, the poachers return to that spot. They raise their car bonnet and throw a tyre on to the ground to make it look like they have broken down. One poacher stays with the car while the other three climb the fence, kill the rhino and cut off its horn. They can then be back in the car and gone in 10 minutes.
Armed with that information, rangers can now patrol the paved roads outside their reserve in the late afternoon to take down licence plate numbers.
When the sun sets, the UAS can begin flying along the perimeter roads, while rangers are stationed in vehicles in strategic places where they can quickly respond to any car that stops at night.
Balule is a small reserve – just 50,000 hectares. In other areas of Africa, the reserves are much larger; poachers often live in villages inside them and strike as far away from civilisation as possible to avoid detection.
Because of the scale of some parks, many poached animals won’t be found for days. In the Kruger National Park itself (which is the size of Israel), the poachers come over the (now unsealed) border with Mozambique to kill animals – away from the public areas of the park.
Fortunately, our models of animal and poacher behaviour can be replicated for anywhere in Africa. All we have to do is import new data for that specific area into the algorithms and generate new strategies to tackle the poachers. Better still, the models are constantly learning and identifying changes in poacher behaviour.
Is it working? Yes. There have been no poachings in Balule in the past eight months. I believe this is largely due to deterrence – the poachers are terrified to go into Balule because the word on the street is that there are machines in the sky that can see at night, and the rangers know where they will walk. This might be voodoo, but it works.
Furthermore, using mathematical modelling has potential to do more than protect Africa’s wildlife. We have recently been asked to adapt our approach to combating illegal fishing, focusing on the east coast of Africa and the Pacific Ocean near Palau in Micronesia.
Here, hundreds of vessels vacuum up the bounty of the ocean. Illegal fishing not only damages the fishing stocks, but also robs maritime nations of hundreds of millions of dollars in badly needed revenue. We are confident that our models developed to target poachers on land will have the same impact on the bandits of the sea.
The London Summit offers the world a great opportunity to use advanced technology to level the playing field against the world’s poachers, whether they be in Africa and Asia or on the high seas. We do not need pompous statements that the world is facing a wildlife crisis; 1,000 dead rhinos in 2013 make this clear. What is needed instead is a commitment to use mathematics and solid science to save the planet’s endangered animals.
By Thomas Snitch, Executive officer, UN Wildlife Enforcement Monitoring System. Dr Thomas Snitch is a Distinguished Senior Professor at the University of Maryland’s Institute for Advanced Computer Studies and the executive officer of the UN’s Wildlife Enforcement Monitoring System at the United Nations University in Tokyo.
Source: Telegraph UK