“...at long last the great moment of development arrives and the plate drops into the solution, as we sit in the dim red light of the dark-room… Yes: a beautiful negative of bamboos arching over where the tiger should be, but alas, no tiger!...But wait a moment. Surely that’s a tiger’s face on the very edge of the negative?...He has seized the hind-quarters from behind and there is his great face, partly out of focus, and just on the edge of the picture.”  

This thrilling glimpse of an elusive tiger, viewed not in the jungle but in a photograph, is from With a Camera in Tiger-Land, the 1927 memoir of the British forester FW Champion about his expeditions in the Shivalik range. Based on a tripwire system devised in the late 19th and early 20th century, Champion’s “flashlight apparatus” involved placing bait to attract the animal and setting up a wire that released the shutter of his reflex camera when the beast tripped it. By creating the ideal conditions for a feline selfie of sorts, the arrangement eventually enabled Champion to produce the first-ever full-fledged photograph of the nocturnal panthera tigris.

Nearly a century on, camera traps are a lot more reliable and sophisticated than Champion’s, generating not just a single blurry image in one night by chance, but potentially tens of sharp and clear pictures per day. This transformation has made these imaging technologies a sound and robust tool of ecological study whose findings contribute towards conservation policies aimed at protecting biodiversity. At the same time, their advantages for research purposes exist alongside troubling ways in which camera technologies in forested areas are used and misused by governments and affect human communities.

In India, the most high profile application of camera traps is probably in the context of national tiger estimation exercises undertaken under the aegis of the Ministry of Environment, Forests & Climate Change. These surveys are carried out with the aim of monitoring a keystone species endangered due to a number of factors – habitat loss to urbanisation and industrialization, prey depletion and a concerningly high rate of poaching, for example. Since the initiation of Project Tiger in 1973, the conservation of Indian tigers has received state support. But it was their local extinction at Rajasthan’s Sariska Tiger Reserve in 2004 that sent shockwaves across the country and the world, prompting the Indian government to set up the National Tiger Conservation Authority, a statutory body under the environment ministry. Since 2006, this agency has partnered with the Wildlife Institute of India and state forest departments to count India’s tigers every four years, combining data from camera-trap-based surveys and sign surveys. The fourth All India Tiger Estimation survey in 2018 was cited in the Guinness Book of World Records as the largest camera trap wildlife survey ever. The fifth one in 2022 was even more ambitious – the full report, titled Status of Tigers: Co-predators & Prey in India, released in late July claims that more than 32,803 cameras were placed at 175 sites for the exercise (compared to around 26,800 in 2018) across 19 states, generating 97,399 tiger photos and identifying 3,080 unique tiger individuals in the process.

Photo for representation only. Credit: DNPWC/NTNC/ZSL Nepal/Panthera.

Image building

The latest iteration of what is colloquially termed the tiger census used a camera-trap-based Capture-Mark-Recapture approach to calculate the abundance and density of India’s national animal through sampling. The report provides an account of the methodology, implemented over three phases. In the first two phases, trained frontline forest workers collected data using a mobile app called M-STRiPES (regarding signs of carnivores and large herbivores, population of hooved animals, presence of humans, facts about vegetation and dung counts) and remotely sensed data provided information about habitat and human impact. In the third phase, the camera traps “were systematically distributed within the sampling area in 2 sq. km. cells…deploying at least one pair of cameras within each cell” spaced at least a kilometre apart. These cameras were installed after the cells had been reconnoitred for the best locations to sight big cats from and sampling was then done concurrently across multiple cells, in blocks of at least 200 square kilometre. With this nested grid division in place, cameras usually operated at each location for 25 to 45 days to obtain images of tigers.

While the survey report is a useful resource for learning about the general methodology of camera-trapping, what does the actual process of setting these traps entail? The Wildlife Institute of India trains local forest watchers, guards and rangers to implement the logistics of the sampling exercise. A Range Forest Officer involved in the 2022 tiger census agreed to describe his work on the condition of anonymity. Explaining the division of the landscape into cells, he said, “We distribute the area into small grids, depending on the equipment, terrain and duration of the survey. After identifying sensitive areas and trails where tigers and carnivores are present with the help of forest watchers and guards, we then deploy the camera for 30-45 days.” Given the necessity of a scientifically useful image, are there any specific ways in which the image is composed? “For tigers, knee-height installation is a thumb rule, to get a good, clear shot.” Installation could depend on the site: “For instance, at a watering hole, each of the two cameras should cover half the span.”

While hardly the occasion for aesthetic considerations, the tiger’s framing by the camera trap is also relevant. “We need pictures of each flank, and adopt a ‘steps-and-stripes’ method wherein we frame the animal according to the distance between hind and forelegs [step] and right and left legs [stripe],” the forest officer said. There are guidelines regarding the equipment itself: “There are two modes – photo and video. We also have to consider the camera’s orientation. To avoid reflection via flash or blurring of the image, cameras should not be facing each other and should instead be tilted at an angle. The time period between photos is a very important aspect, whether 15, 20 or 30 seconds. We aim to get at least one or two clear photographs during a tiger’s crossing [between the cameras].”

Photo for representation only. Credit: Timothy A Gonsalves/Wikimedia Commons [Creative Commons Attribution-Share Alike 4.0 International License].

Once the images have been snapped and stored in the camera’s memory card, the frequency of collection varies, depending on the accessibility of the site, the purpose of the data and the capacity of the card. In case the location has vehicular access, collection could take place within a day or two, but pedestrian access takes longer. The forest officer added, “In some cases, such as an animal conflict or an injury requiring medical intervention, we might need continuous photos which deliver daily captures but the regular cycle is 15-30 days.” The frontline staff then dispatches the collected images through a chain that goes from the Range Office to the Division Office, then the state-level coordination committee, before finally arriving at the Wildlife Institute of India for analysis by its scientists, and storage in a national repository of camera trap photographs of tigers and leopards under the direct control of National Tiger Conservation Authority.

Scientists from the Tiger Cell at the Wildlife Institute of India are repositories of insights into the technical aspects of the camera traps’ operation, the processing of the images and the implications of the camera-trapping method for research. YV Jhala, the former dean of the Wildlife Institute of India who led the tiger census in 2022, points out that an important goal in camera trap manufacturing “is to improve the camera image and make the technology simple, because currently it’s very expensive and requires a lot of effort”. Historicising the photographic hardware used in the national tiger estimation exercises, he said, “They used to be based on infra-red light, but now come with in-built heat and motion sensors, which was the major breakthrough in technology.” How do scientists decide which equipment to use? “There are many brands which vary in performance and pricing,” noted Jhala. “We go for the middle range ones – they provide the appropriate resolution (5 megapixels or more). More than resolution, the most important factor is how fast the camera is…it is the synchrony between flash and shutter that matters for the purpose of our work.”

As far as the camera arrangement itself, said Jhala, “We first look for tracks [to place the cameras]. Lots of groundwork is required – rake marks and scat marks are used to identify the trails. We have standardised the height at which cameras are installed, the objective being perpendicular shots from a distance – with full frame pictures of the flanks from both sides.” Ayan Sadhu, a contributor to the 2022 tiger census report, elaborated further on the apparatus: “These are simple cameras paired with digital sensors that sense both temperature and movement simultaneously. Whenever a warm-blooded animal – that is, one whose temperature the heat sensor can detect to be greater than ambient temperature – crosses between cameras, the motion sensor is activated.”

Photo finish

Heat and motion sensors are relatively new additions to camera traps. Jhala’s colleague at the Wildlife Institute of India, Qamar Qureshi, who has been part of the tiger census since its inception, talks through the various methods of tiger counting in independent India. “In the 1970s,” said Qureshi, “Saroj Raj Chaudhary [then Forest Conservator for Orissa] devised the method of counting tigers on the basis of pugmarks. The pugmarks would be traced on a glass plate and then transferred onto butter paper or cast into plaster. This was a method ripe for human error and not scientifically sound.”

Photo for representation only. Credit: Brian Gratwicke/Wikimedia Commons [Creative Commons Attribution 2.0 Generic License]

In India, the use of camera traps to estimate populations was pioneered in the 1990s by Ullas Karanth, who relied on the classical Capture-Recapture model to calculate the number of tigers in a given area. Qureshi continued, “With the establishment of the National Tiger Conservation Authority in the mid-2000s, that model has been replaced by Spatially Explicit Capture Recapture based on maximum likelihood. It is the most robust method used worldwide.” He takes a step back to offer an overview of camera trap technology’s evolution through the census cycles: “When we started, we focused on identifying individual tigers through their stripes, then their ranging patterns, and now we can use camera traps to study non-patterned animals too.” While acknowledging that the most precise method of tracking animals is satellite telemetry, Qureshi feels that a 2 square kilometre camera trap density does allow for significant success. “We want inference about a given area to be drawn across time scales/periods and this method allows us to do a 1:1 comparison across time periods.”

Ullas Karanth, the tiger ecologist who introduced camera traps to India, has been sceptical of the tiger estimation surveys and is among a group of scientists who has engaged in debate over their statistical methodology and extrapolation. He was equally critical of them in my interview with him: “Lots of people use traps in a random, non-optimal way. Thousands of camera traps have been bought by the government but most people do not understand the problem with tiger surveys.”

Beginning with Bill Goodson’s Trailmaster cameras as a graduate student in Florida in the 1980s, Karanth pursued his interest in wildlife sampling by working with the Capture-Recapture model. Since then, he says, paradigm shifts have occurred on both the technological and conceptual level, from the transition from film rolls to digital memory cards, to the optimisation of camera trap placement to the refinement of statistical modelling underlying the analysis of images. However, he cautioned, “None of the developments get rid of the need for knowledge of natural history and animal ecology.” His chief complaint is that the government has not shared relevant data in the public domain for independent scientists to analyse and verify: “Unfortunately in India the official management of forests is not run by scientists – we get treated like technicians.”

The 120 scientists at the Wildlife Institute of India’s Tiger Cell have a laboratory equipped with workstations having the computational capacity to process the amount of data received. Sadhu shares that, similar to the collection, there are multiple phases of sorting and analysing it. First of all, GPS locations are checked and photographs geotagged or paired with accurate locations – “a practice that is hugely important for tracking poaching by matching recovered skins with those of individual tigers in our database.” Next follows the artificial intelligence-based segregation of images into various categories and identification of individual tigers. Finally, researchers manually check the data for errors and verify it.

Photo for representation only. Credit: Ali Arsh/Wikimedia Commons [Creative Commons Attribution 2.0 Generic License].

Another scientist contributor to the 2022 tiger estimation survey, Shikha Bisht speaks about the way in which the vast corpus of camera trap images are collated into scientifically useful sets. Over WhatsApp, she said, “All the nearly four crore animal images are compiled. We use a geotagging software called the Camera Trap Data Repository Analysis tool (CatRat)... CatRat stamps camera trap images with GPS locations and prepares the data for further processing. After this, an AI-based auto-species segregation tool… is used to segregate these camera trap images into species. Tiger and leopard images are then processed through the ExtractCompare program to get an individual tiger’s number.” However, despite the efficiency of artificial intelligence tools, “with sensitive data like tiger and leopard images, the manual intervention of trained biologists is imperative. Unless the AI tools can process images with 100% accuracy, manual intervention will still be required.”

From these images, a number of conclusions can be drawn. “We observe behaviour, interactions between animals, foraging behaviour (waking, resting etc.), tiger-human conflict, and so on,” said Sandhu. Over the past 17 years, tiger population dynamics have been extensively studied using camera traps – abundance, density, richness, activity, dispersal, life history and interactions with other predators are some of the parameters on which studies have been conducted. The other prominent utility of camera trap images that ecologists reiterated was their use in helping curb illegal wildlife trade – by identifying poachers, aiding their prosecution or by facilitating the matching of photos of tiger skins to those of the individual tigers in the central database, marked by their unique stripes.

As scientific instruments, camera traps provide crucial data and information about animal populations, behaviour and forest habitats. At the same time, the form’s ancestral links to colonial regimes of hunting and policing and contemporary links to what has been termed the “militarization of conservation” cannot be dismissed. Karanth did acknowledge the camera traps’ affinity with other technologies of surveillance such as CCTVs and drones (now flying in forests), arguing that “this is not a problem of camera traps alone.” It is, however, a problem. Conservation geographer/sociologist Trishant Simlai delivered a talk in 2021 based on his dissertation Negotiating the Panoptic Gaze: People, Power and Conservation Surveillance in the Corbett Tiger Reserve, India in which he critiqued conservation surveillance technologies including camera traps. His ethnography-based research at Corbett revealed how camera traps extended patriarchy into women’s spaces in the forest, entrenched caste power through scopic violence and criminalised scheduled tribes, and indigenous communities by weaponising the categories encompassing illegal wildlife activities against them. The ensemble of drones, camera traps and electronic eyes constituting “[c]onservation surveillance, when used without ethical safeguards and with complete impunity by the state, can lead to control in the most repressive ways.”

Given, as Simlai found, that images of humans engaged in perfectly legal activities can be captured and circulated, what are camera traps’ implications for privacy? Bisht said the Wildlife Institute of India receives images of tourists, guards, researchers and villagers, used “to assess relative abundance index of tourists and [the] presence of humans”. This data is stored separately in hard drives in the central repository, sans time limits and privacy policies, she claimed. It is, she said, the property of the “National Tiger Conservation Authority and state forest departments. They decide whom to share it with.” The National Tiger Conservation Authority did not respond to a request for comment.

It is ironic that the way in which humans seem to enter the frame of the forest camera is often non-consensual. An independent scientist who wished to remain anonymous remarked, “No doubt the visuals in front of the camera are extremely important for science and for the preservation of our ecosystem. However, the humans off-camera must not be rendered invisible. What is the status of livelihood, health and education in forest communities that may be driving them to illegal activities? Everybody talks about human impact on wildlife, but what is the impact of wildlife technologies on humans historically living in or off forests?” The scientist reiterates that their experiences as a researcher within the Indian government’s wildlife sector alerted them to the importance of involving all stakeholders – scientists, legislators and, vitally, local and native communities – in developing policies. Camera traps are a powerful mode of advancing human knowledge of other species and protecting our shared environments. They must be deployed with judiciousness on the part of governments and researchers who recognise that the values governing care for species imperilled by humans apply to vulnerable humans too. Else, just like Champion’s tiger, human forest dwellers will be left “partly out of focus, and just on the edge of the picture”.

Kamayani Sharma is an independent writer, researcher and podcaster based in New Delhi. She is a Kalpalata Fellow in Visual Culture Writing for 2022.