It’s nine in the morning. The scent of elephants hangs heavy in the air. Every time a gust of wind rustles the leaves, the shadows cast by the summer sun dance on the forest trail. But the scene unfolding on a patch of green in Tamil Nadu’s Anamalai Tiger Reserve (ATR) verges on the comic: forest watcher Karthik, in his dark green camouflage uniform, is crawling on all fours.
Three metres away, forest guard R. Ayyappan bends beside a tree. He is monitoring a small camera strapped to the tree trunk three feet from the ground. The camera faces the forest trail. As Karthik clambers into its view, we wait for a red light to blink. That would be the sign that the special heat sensing and movement sensing remotely activated camera is ready to take pictures — of people, deer, or tigers.
“No, the light isn’t coming on,” says Karthik. “We need to lower the camera some more to ‘catch’ the tigers.” Both Karthik and Ayyappan adjust the forest green metal box that houses the camera. They are on their weekly routine of checking the eight pairs of cameras installed in their 358-hectare beat (a park’s smallest administrative unit) in the ATR.
This year, almost 15,000 pairs of camera ‘traps’ are being placed across protected areas and reserve forests in 18 States — all to count the elusive national animal. Photographs of the big cats and indirect tiger signs will contribute to the fourth All India Tiger Estimation, undertaken to ascertain India’s current tiger population. But scientists will arrive at the final number only by the first quarter of 2019, for the process is lengthy, the science complex, and the implementation of the exercise challenging.
Signs, cameras and science
“India’s tiger census is easily the largest such effort of its kind across the globe,” says Debabrata Swain, former Member Secretary of the National Tiger Conservation Authority (NTCA).
The quadrennial survey will cover almost 4,00,000 sq km of tiger habitats in India. This time, the almost 500-day-long process began in December 2017 across different parts of the country. Designed by scientists at the Wildlife Institute of India (WII), it involves four distinct phases. In the first phase, forest department personnel such as Ayyappan and Karthik tramped through the tiger habitats in their beats for a week, recording signs of large cats (such as scats or pugmarks). This provides information on the number of tiger signs seen every kilometre, which is later used to supplement camera trap data and predict tiger numbers in larger areas where such cameras cannot be used.
The field teams also note signs of prey and the area’s dominant vegetation. With many tiger habitats bordering villages and towns, human disturbance in the region is noted too, for it can play a role in tiger distribution. In the ATR, for instance, tea and coffee plantations completely surround some of the reserve’s lush evergreen forest patches.
“Estate workers have seen tigers walk this way sometimes,” says Karthik, pointing to the wide forest trail that skirts a coffee plantation further ahead. “We saw tiger scat here during the sign survey too.”
During the sign survey, Karthik, Ayyappan and other forest guards in the ATR identified numerous such locations where camera traps had the maximum chances of obtaining tiger photographs. From these, ATR’s in-house biologist Arumugam Rathinasamy shortlisted around 480 locations, distributed across the reserve’s three ‘blocks’ of roughly 300 sq km each. He had to take into account factors such as the distance between camera traps, which plays a crucial role in obtaining precise tiger numbers.
Using Global Positioning Systems (GPS) to locate each of these spots, forest guards and watchers installed 170 pairs of camera traps in the first block by late December and monitored them for 25 days, setting the scene for Phase III. (Phase II involves WII’s scientists accessing remotely sensed landscape and habitat information for later use in the analysis.)
“Direct tiger sightings are very rare here,” says Ayyappan, as he checks whether the camera’s batteries are still working. “Though we can infer tiger presence from pugmarks or scats, camera traps give us undeniable proof that there are tigers in our beat.”
Analysing the numbers
The photographs, however, are not to confirm tiger presence alone. Each tiger has a unique stripe pattern. By comparing them, either visually or using a software, scientists can identify individual tigers. The total number of tigers caught on such camera traps is the minimum number of tigers in the region. In the ATR (which is a medium-density tiger landscape), the last census showed a minimum count of 20. This number, however, is not an estimate of the tiger population in the area. Cameras may not capture every single tiger either because some of them may not have walked that way or the cameras may have malfunctioned when they did.
To account for these eventualities, and also for methodological reasons, scientists study the number of times a specific tiger was captured on camera, as well as the number of times the cameras recaptured the same tiger again. This method, which also factors in the locations of the camera traps to estimate the population size, is known as ‘spatially explicit capture recapture’ (SECR).
Once camera-trapping wraps up across the country, all the teams will compile their data and send it to the WII and the NTCA for scientific data analysis. Then, in a process known as index calibration, the tiger numbers obtained from the more accurate camera-trapping exercises in different reserve forests will be integrated with coarser information from sign surveys and other data that cover larger areas. This will provide an estimate for the entire country.
Though the survey this time follows roughly the same methods as before, there are three major changes, says Vaibhav C. Mathur, Assistant Inspector General, NTCA. First, a majority of the States and reserves will use M-STrIPES, a mobile-based application, to collect data on the field. (The ATR, however, is not using it this time).
Second, to obtain more precise estimates of the tiger numbers, the area in which a single pair of camera traps is deployed (called a ‘grid’, usually measuring four sq km) has been decreased to two sq km. This means that more cameras are being used this time, making the current survey more intensive.
“With this, we also hope to obtain information about smaller fauna through the same camera traps so that we can work out which areas they occupy,” he says. “It would be possible to see how their ranges overlap with those of the tiger, which can strengthen the ecological tenet of using tigers as an umbrella species. Estimates of smaller fauna may require further refinement, which can be tried in the future.”
And the third change, he adds, is that for the first time ever, India will be conducting the census along with the three other tiger-range countries — Nepal, Bhutan and Bangladesh. Representatives from these countries have completed their training in India.
“We have come a long way from Project Tiger to the NTCA,” says V. Ganesh, Chief Conservator of Forests and Field Director of ATR, commenting on the changes in tiger monitoring and conservation over the decades.
The pugmark census
The first method used to count India’s tigers, the pugmark census, was already in use during the launch of Project Tiger in 1974. Once every four years, forest guards and watchers would walk across tiger habitats over two weeks, identify tiger pugmarks, and take their plaster casts. Based on the shape and other measurements, these pugmarks would be assigned to individual tigers to arrive at an absolute count.
But this meant obtaining every single tiger’s pugmarks (the left hind paw, specifically). It also assumed that the thousands of forest personnel making pugmark casts on the field were equally skilled at the job, and that a tiger’s tracks looked the same regardless of whether it walked on soft mud or firm ground. “There used to be a lot of confusion in identifying pugmarks,” admits Ayyappan, who has been in the service for 20 years now and has participated in a pugmark census.
By the late 1980s, Indian scientists began to question the obvious limitations of a pugmark census and started to explore other methods of counting the endangered species in the wild. F.W. Champion of the Indian Forest Service had already taken the first photograph of a wild tiger in the 1920s, using a rudimentary trip-wire enabled camera trap. The world over, scientists began discovering that individuals of several species — from mountain lions in Panama to bushbuck in South Africa — could be differentiated by the markings on their coat.
In the early 1990s, in Karnataka’s Nagarhole National Park, tiger biologist Ullas Karanth used automated camera traps for the first time to individually identify tigers and estimate their numbers. Scientists began using this method to estimate the numbers of tigers in Sumatra, jaguars in Bolivia, and leopards in Africa. Finally, in a major yet reluctant transition, the first camera traps found their way into India’s tiger estimation exercise in 2006.
However, camera-trapping to estimate animal numbers is a tool that will work only if the exercise is conducted with absolute meticulousness. Carelessness — be it an incorrect placement of cameras or sloppy data analysis — can inflate or underestimate India’s tiger numbers, a statistic followed closely by wildlife enthusiasts across the world.
Do the numbers add up?
According to the WII and NTCA, India’s tiger population has been observed to increase at a rate of around 5.8% per year since 2006. The estimation in 2014 pegged tiger numbers at 2,226. The same year, the news that global tiger numbers grew from 3,200 in 2010 to 3,890 in 2014 caused much cheer.
But when conservation body Panthera Foundation’s Abishek Harihar and his colleagues studied the camera-trapping methods used in India and Nepal in 2010 and 2014, they claimed that the increase in the tiger estimates could have been just a by-product of changes in the survey methods, such as deployment of more camera traps and changes in analytical methods. They found that in India alone, the increase in tiger numbers corresponded with a 538% increase in the number of camera traps deployed, resulting in a 144% increase in the number of tigers photographed. The 2014 surveys were conducted in 32 additional locations, which in itself could have contributed to the rise in numbers, they said.
Though officials claim that the total area of approximately 4 lakh sq km that will be camera-trapped for the current census will not change, the availability of more camera traps could result in new locations being added to the exercise this time, too. For instance, the ATR (which has nearly 75 more camera traps this season) has finished its camera-trapping exercise and lent its camera traps to five surrounding forest divisions — Theni, Dindigul, Srivilliputhur, and the Kodaikanal Wildlife Sanctuary— where camera-trapping will be conducted for the first time. This means that for the current census, about 3,000 sq km more is being camera-trapped around the ATR alone than was done in 2014.
“If more new areas are included as part of the current survey, it would be important to ensure that valid comparisons are made in terms of the sampling frame, analytical methods, and other issues that we raised in our paper late last year,” says Harihar. Government officials, however, claim to have incorporated all these aspects into their inferences.
That’s not true, says Karanth, who has been critical of the methods and analysis used. The present method of analysis does not address the inherently huge variations in the relationship between the chances of capturing tigers on camera and seeing its signs on the field. This renders the index-calibration method — which is how reliable estimates of tiger numbers derived from small areas are extrapolated to wider regions — invalid, he says. “The results of the national surveys of 2006, 2010 and 2014 do not make biological sense in the light of what we now know about how tiger populations function.”
Identifying this as a potential problem in obtaining accurate tiger estimates, Karanth and his colleagues examined the statistical models that generated these tiger numbers in a study in 2015.
“Our study found that integrating the two methods — which is what the index calibration exercise of the current census tries to do — is often a futile exercise that could be generating inaccurate tiger estimates for India,” says Arjun Gopalaswamy, its lead author and visiting scientist at Indian Statistical Institute’s (ISI) Bengaluru Centre. “Based on this mathematical finding and India’s tiger estimation results, the urgent and simplest first step is to re-analyse India’s tiger census data of the last three surveys to figure out whether India’s tiger numbers are rising or declining. Repeating the same mistakes will not help,” adds Gopalaswamy.
Refining the method
Just as the pugmark census was replaced with a more robust camera trap system, there is an urgent need to update existing data analysis methods with new techniques to make the exercise more precise and reliable, say scientists.
Another study published last year by scientists at the ISI developed a new model using a refined statistical technique that helps to better integrate the data obtained from the two different methods — camera-trapping and sign surveys — to count tigers on a large scale. If implemented, this could reduce the inaccuracies in India’s tiger estimates. Other countries are already adopting some of these new approaches and moving away from index-based approaches. For instance, the central African nation of Gabon is adopting some of the new capture-recapture methods to count its forest elephants, says Gopalaswamy, who is on the Gabon Technical Unit formed to implement this census technique. “Uganda is also encouraging new approaches for its lion census,” he adds. Existing methods would also need to be implemented more strictly, feel others. The methods described in detail in the tiger estimation’s Phase IV protocol (which entails annual camera-trapping to ensure that important tiger populations, such as those in tiger reserves, are monitored continuously) are crucial to capture the huge natural variations in tiger population densities, says Karanth. “However, this protocol has been merged and diluted with Phase III in the current survey.”
If implemented properly, the intensive Phase IV surveys can give a lot more information than just numbers, says Gopalaswamy. While it could mean more work for ground personnel like Ayyappan and Karthik in the ATR, they would be able to generate better data with these surveys.
“Obtaining vague total numbers isn’t too useful scientifically. It’s the other information that we get — births, deaths, movement, sex ratios — that tell us more about tiger population dynamics,” says Gopalaswamy. According to Karanth, conservationists should not be satisfied with “substandard methods”, given the crores of rupees and the massive effort that go into the census.
Back in the Anamalais though, Karthik and Ayyapan are oblivious to these debates. The camera-trapping season in their reserve has come to an end. All the additional work that entails a ‘tiger census year’ is done for now. They are back to the duties that keep them busy round the clock — from fighting the occasional forest fire to patrolling the forest trails every single day.