One day in December 2013, Shivamallappa Basappa, a farmer from Chikkabaragi village in Karnataka, was returning home with his cows when a tiger attacked him and dragged him into the forest.
A few hours later, by the forest’s edge, his son found all that remained of him—a skull and remains of a leg.
When other farmers from their village heard what happened, they ransacked a forest lodge, set alight two vehicles belonging to the forest department, and called for Basappa’s death to be avenged by tracking down the creature, dousing it in gasoline, and setting it aflame.
A recent paper in the conservation journal Oryx describes what happened next.
Before the aggrieved villagers could act on their wishes, forest officials used a recently-developed software to correctly identify the man-eater in the forest, which turned out to be a bedraggled male, and dispatched it to a nearby zoo.
This open-source software was created by wildlife biologist and lead author Ullas Karanth six years ago.
After an abortive foray into engineering, Karanth trained as a wildlife ecologist at the University of Florida in the late ‘80s and at Mangalore University. Over the past 20 years, he has conducted pioneering research on tigers and their prey at Nagarhole National Park, and is recognized for his conservation efforts in the region. He worked on the technology with Lex Hiby, a population biologist who had first designed the software to help tell grey seals apart around the British Isles.
How it works
The software ExtractCompare, the first of its kind for tigers, projects photographs of tigers on to a three-dimensional surface, and tallies their stripe patterns, distinct as a barcode, against images stored in a database. When a “high similarity” score is reached, a successful identification is made.
“This rapid pattern matching system speeds up the matching process when thousands of comparisons are involved,” said Karanth. But no software, no matter how sophisticated, has yet topped a field researcher’s well-trained eye. “The final match is still done visually,” added Karanth.
The speed and accuracy of the software has revolutionized tiger research, wildlife management, and particularly conservation efforts, by improving census-taking and the monitoring of wildlife crime. With the results of the 2014 tiger census being announced next month, both of these are crucial concerns.
How many tigers do we have?
On Nov. 26, India’s environment minister Prakash Javadekar announced that 274 tigers had died in the past four years, with less than half dying of natural causes. As Down to Earth magazine pointed out, almost half of these tigers died outside core reserve areas, which are likely to degrade even more, following the government’s steady weakening of forest protection laws that facilitate quicker clearances for project developers in the eco-sensitive zones around tiger reserves.
In the face of intensifying pressures on their habitat, getting the number and distribution of tigers right is more important than ever. After the last tiger count in 2010, which put the tiger population at 1,706, the government-run Project Tiger began to use Karanth’s software, with a central database of camera trap images of tigers to better monitor their populations in reserves.
This was a much-needed upgrade from the widely discredited “pugmark method” which the government employed for tiger censuses until 2005. This method, developed by an Indian forester in 1966, involved the comically difficult task of scouring the country’s potential tiger habitat (300,000 square kilometers in all, and roughly the size of Italy) for tracings or casts of the left hind prints of individual tigers. Surveyors were overzealous—one even claimed he’d traced a wet paw print as it faded on a rock—and they unfailingly turned in inflated counts.
In 2007, after the introduction of a better assessment technique using camera traps (also developed by Karanth), the tiger population of India plummeted from 3,624 tigers, as the 2002 census claimed, to 1,411—less than half that number.
At least a good number of those disappearing tigers were imaginary. In the seven years since, poachers who trade in tiger parts have killed many hundreds of animals, and reliable estimates of these killings are hard to come by. This year, there have been five confirmed instances of poaching, with 41 under investigation, while last year saw 15 poaching cases, aside from 44 still “under scrutiny.”
But the software has improved these tallies over the past four years, according to Abishek Harihar, a Bangalore-based tiger conservationist. “When skins have been seized in Delhi passing through to, say, China, this software has helped ID where the individual was poached,” said Harihar. “It has more than halved the time taken for identifications.”
Spotting the right individual in the national camera trap database, which holds more than 1400 tiger individuals, could take weeks. Samba Kumar, a biologist who co-authored the Oryx study with Karanth, says that tallying an individual against the images in their 750 tiger-strong regional database without the software would take 10 to 15 days. “On the contrary, with the help of ExtractCompare software, each image will take less than 30 minutes to establish its identity,” he said.
It has also provided park managers with better information on which to base their decisions. “Not all tigers that enter into conflict with people are man-eaters or habitual cattle lifters,” said Bivash Pandav, a conservationist who studies the tigers of the grasslands and forests of the Eastern Himalayas.
He attributed a large part of the human-tiger conflict to an ill-starred combination of tigers’ social organization, which forces young males out of the forest to establish their own territory at the age of two or three, and the increasingly degraded patches of land that link up the core forests where tigers live.
“By clearly establishing the identity of the individual, which this software helps you do within 24 hours, you can quickly tell if it’s a stray instance or a habitual offender that should be removed from the area,” Pandav said. “It’s a powerful tool in managing conflict.”
Powerful tools don’t always lead to happy results.
Kumar asserted that tragic cases like Basappa’s were outliers, and that any uptick in human-tiger conflict in his fieldsite was due to high reproduction rates—the happy consequence of tigers finding themselves in a secluded, prey-rich environment. Most instances of conflict, he said, involved young male tigers dispersing out of these tiger-saturated spots in search of a new territory to call home; their searches taking them through hamlets where a stray cow gave them cause to tarry. If biologists had their way, all these blameless individuals—once identified by CompareExtract —would be relocated to a prey-rich patch of forest to proliferate in peace.
But forest officials, often acting in the face of public ire, hastily catch such tigers and consign them to zoos. “Ultimately,” said Kumar, “the decision is theirs.”