We analyzed tiger habitat annually, on January 1^st of each year from 2001 – 2020, within the “likely resident indigenous range” of the tiger as defined in Sanderson et al. (in press) ( Figure 2A ). The indigenous range mapping indicates which ecoregions, or parts of ecoregions, were most likely to have resident tigers before human beings became a significant factor shaping the distribution of the species (Sanderson et al., 2019). This definition is consistent with how the indigenous range is defined by the IUCN Green Status Assessment process (Akçakaya et al., 2018; see also Stephenson et al., 2019). Within the study area, raster analyses were made over the 300 m grid cells coincident with the ESA CCI Landcover data (European Space Agency, 2017) used for the structural habitat mapping (see below). Previous TCL mappings used a base resolution of 1 km cells, so our landscapes are approximately 3x better resolved than earlier efforts.

We divided the study area into four zones along ecoregional boundaries, representing major differences in tiger biology and status: South Asia, Southeast Asia, Northeast Asia, and Central Asia ( Figure 2A ). These zones were used to create a grid-based system for mapping species observations as areas of analysis for the human footprint (Sanderson et al., 2022) thresholds, and for assigning minimum patch sizes, as described below.

We also developed maps of ecoregions or parts of ecoregions where tigers had been declared extirpated, including broad swaths of the Black Sea region, Central Asia, Indochina, and central China, for each year from 2001 – 2020 (Sanderson et al., in press). For extirpated areas, we assigned the approximate date of extirpation and provided supporting references.

We systematically reviewed the literature to identify peer-reviewed studies, government and non-governmental organization reports, and in some cases, newspaper accounts to develop a temporally and spatially explicit observational database from 1995 – 2021 for tigers. Observations include any information of direct observations of the species, including systematic surveys, that document whether or not the species was detected. We also included surveys for other mammal species (e.g. small cats, wild cattle) using techniques that would likely have detected and reported tigers, especially in parts of the range and/or time periods where or when no tiger-specific studies were available.

We began by searching Google Scholar and Web of Science using keywords “tiger(s)” OR “Panthera tigris” and “survey” OR “field” in journals focused on ecology, environmental science, and conservation. Next, we filtered the extensive literature by reviewing titles and abstracts and, where necessary, article contents, keeping only papers that provided first-hand observational data on tigers or where second-hand summaries provided access to otherwise unpublished information. We also contacted tiger biologists and conservation officers working in tiger range countries to share information and identify sources our review might have missed, including non-English language sources and experts ( Table S1 ). Finally, for a few localities where scientific accounts were unavailable because of censorship or lack of data sharing, we supplemented scientific observations with accounts of human-wildlife conflict drawn from a search of newspaper and magazine accounts, confirming observations through independent sources and/or photographic evidence. We maintained all the bibliographic details in a read-only, public bibliography: https://www.zotero.org/groups/2516009/tiger_conservation/library. A full list of sources is provided in Table S2 .

The four types of data were:

Observations were located by coordinates, if available, or by matching to a standardized grid reference, if not. We created an overlay of grid cells that varied in size in different zones of tiger range, depending on a summary of tiger home range information ( Table S3 ). In South Asia, we used cells with a grid side of 4 km (area of 16 km²); in Southeast Asia, including Sumatra, we used 16 km cells (area of 256 km²); and in Northeast and Central Asia, we used 32 km cells (area of 1024 km²). If precise coordinate data were unavailable in the paper or from the study authors, we georeferenced map figures and extracted spatial locations according to the overlying grid cells.

Observations were dated to the day of the observation when possible, and if not to an observation period, with a start and end date. If survey dates were given in months (e.g. March 2004 – October 2004), we used the first and last days of the corresponding months. Where observations were in years, we used the first and last days of the corresponding years.

All observations were tagged with the observer’s name, institution, email address, and reference. For multi-authored studies, we identified the observer as the survey leader, where given, or the corresponding author’s name, where not. We agreed, as a matter of security for the safety of tigers and to respect the research interests of the investigators, not to reshare the point locality information used for the analysis. They do not appear on the website or in this paper.

As ambush hunters, tigers need vegetative cover to stalk and kill prey (Sunarto et al., 2012). They also use thick vegetation to hide their cubs during weaning (Smith et al., 1998). We used land use/land cover data developed by the European Space Agency to identify land cover classes that are possibly tiger habitat and vegetation height product derived from NASA’s Landsat sensor (Potapov et al., 2019). Both datasets were rescaled to a common 300 m grid. For tigers, structural habitat was defined land use/land cover classes where tigers are typically found – essential habitat types with sufficient cover for tigers to hunt and raise cubs ( Table S4 ). For mixed mosaic classes, we add another criterion, including areas with at least 5 meters of vegetation height ( Figure 2B ). For the Northeast Asia zone, we excluded areas 1000 meters above sea level; elsewhere we applied an elevation threshold of 3350 meters, though note that rarely tigers have been observed at higher elevations in the Himalayas (Adhikarimayum and Gopi, 2018; Shrestha et al., 2021).

Factors invisible to satellite sensors impact where tigers can live (Redford, 1992; Harrison, 2011). Tigers are illegally hunted or killed in many parts of the range after human/wildlife conflict (Nyhus and Tilson, 2004; Karanth and Gopal, 2005; Musavi et al., 2006; Lubis et al., 2020). Tigers may also be depleted by the lack of prey caused by indirect competition with human hunters (Miquelle et al., 1999b; Karanth et al., 2004; Johnson et al., 2006). Because we lack range-wide, spatially and temporally explicit data on prey depletion and human disturbance, we use a proxy: the human influence index, or informally, the human footprint (Sanderson et al., 2002; Venter et al., 2016; Sanderson et al., 2022). The human influence index is a unitless, weighted sum based on human population density, infrastructure, accessibility, and power consumption on a 0 – 60 scale ( Figure 2C ). The result is a gradient map of human impacts that have been widely correlated to human disturbances and range collapse in large mammals (Yackulic et al., 2011; Di Marco et al., 2018; Tucker et al., 2018).

We recorded the human influence index values and zones where tigers have been observed between 2001 – 2020. Following methods developed by Sanderson et al. (2006), we randomly sampled the human influence index at the same number of locations within each zone. We repeated this analysis a thousand times for each year for each zone. Then, we calculated frequency histograms of human influence index values for the tiger detections and the random sample for each zone and subtracted them to show human influence values where tigers are more or less likely than random selections would suggest, equivalent to a chi-squared test (Nikulin, 1973). The lowest human influence value where the difference in frequency between the tiger and random distributions is zero, we call the “social tolerance threshold” ( Figure S1 ).

To define effective potential habitat, we excluded areas of structural habitat where the human influence index value was greater than the social tolerance threshold for a given zone and year ( Figure 2D ). We applied the threshold from the Northeast Asia zone to the Central Asia zone.

We applied a two-step delineation process to group patches of effective potential habitat into landscapes. First, we found all patches of contiguous cells and tested them against core patch size specific to ecoregions ( Table S3 ). A “core” patch size is defined as an area large enough to maintain at least five tigers, based on densities observed in the literature. “Stepping stone” patches were also recognized that were one-tenth the size of a core patch. These small patches are considered to provide temporary respites during movement, but the patches are not large enough to maintain a tiger for a breeding season. In a second step, we tested how patches were close enough to be considered potentially connected. If core patches lay within 4 km of each other, or were connected with “stepping stone” patches within 4 km, they were delineated as part of the same landscape ( Figure 2E ).

We attributed landscapes, or portions of landscapes, into one of six classes based on size, the presence of the species, and the survey effort applied to detect them, based on the observational data for the last five years. The classes are summarized in Table 1 . We analyzed landscapes using underlying state/province boundaries of countries (administration level 1 as defined in the Global Administrative Database – see Anonymous, 2021). States/provinces represent potentially important political differences in management regime and cultural situations that might differentially impact portions of species conservation landscape ( Figure 2E ). Observations with coordinates were assigned with a point in polygon analysis. Observations located by grid cells were assigned based on the centroids of the grid cells to the overlapping landscape.

To integrate various survey data, estimate the effective amount of survey effort and account for variation in the size of landscape patches, we fit a multi-scale occupancy model (Nichols et al., 2008), in which the landscape patches were the coarser scale and grid cells were the finer scale (and whose dimensions were defined approximately based on average tiger home range size in that zone). In multi-scale occupancy models, occupancy at the finer scale is conditional on occupancy at the coarser scale – in other words, tigers can only occupy a grid cell within a landscape patch if the landscape patch is occupied; however, there can also be unoccupied grid cells within an occupied landscape patch. In this context, the finer scale occupancy would be equivalent to the probability of use. To fit the multiscale occupancy model, observations and effort associated with sign surveys and camera trap surveys with full details were attributed to grid cells nested within landscape patches. Occupancy at the coarser scale was a function of random effects for state/provinces, and fixed effects for area of state/province and proportion of the area protected. We standardized area of each subnational unit before fitting models. We used constant (e.g., no covariates) probability of detection for camera trap surveys and constant probability of detection for sign surveys.

Models were fit in JAGS (Plummer, 2003) run from Python using a Python interface to JAGS, PyJAGS (pyjags version 1.3.8) to analyze Bayesian hierarchical models. We chose a Bayesian framework because it allowed us incorporate the random effects detailed above more easily.

For each landscape patch, we estimated the unconditional probability that tigers were present in a landscape (given the landscape patch’s potential effective habitat area, protected area status, state/province, and ecoregion). We also estimated the conditional probably that tigers were present given the observational data from the last five years (e.g. for the 2020 analysis, observational data applicable, based on the start and end dates of the study, from 2016 – 2020). If ad hoc or camera trap summary data indicated that a landscape patch was occupied and its conditional probability was less than 1 (i.e., no tigers were seen in the detailed camera trap data or sign survey data), the conditional probability was changed after model fitting to 1. We measured survey effort of landscape patches by taking the ratio of the conditional to the unconditional probability subtracted from 1.

Tiger conservation landscapes were defined as areas with a conditional probability of the presence of tigers of >=99%. Survey landscapes have a lower conditional probability (<99%) and a survey effort below 0.6. Restoration landscapes have a lower conditional probability and a survey effort greater than 0.6.

After classifying the landscapes into the six classes, contiguous landscapes of the same class were recombined to create the final landscape delineation, as shown in Figures 2F ,  3 .

We created several automated summaries to make the information contained in these data more tractable to policymakers and on-the-ground managers ( Figures 4 , 5 ; Tables 2 – 4 ). Each landscape is analyzed against the World Database of Protected Areas (UNEP-WCMC and IUCN, 2022) to measure the extent of protection and the Key Biodiversity Areas (IUCN, 2016) to measure its importance for meeting Sustainable Development Goals (United Nations, 2015) and other biodiversity-related targets. This information is available for each country via the project website at act-green.org

The steps above were implemented on Google Earth Engine (Gorelick et al., 2017) in a task-based architecture, implemented in Python (Python Software Foundation, 2022) and designed to be recalculated as often as the underlying data change. All areas are reported based on Google Earth Engine’s area calculation procedure, which treats each pixel element as its true three-dimensional area, not dependent on any particular map projection. The open-source code is available at https://github.com/SpeciesConservationLandscapes.

Each tiger observation study was reformatted into a data entry form compatible with the database schema. Another analyst (usually the senior author) separately reviewed each data entry form before inclusion in the database. Through extensive online consultations, we also reviewed the landscape outputs with tiger biologists and conservationists noted in Table S1 . Our results are replicable by other scientists as our code is open-source, and we identify in detail the sources used ( Table 2 ). Sensitivity testing of the model was conducted, as described in Sanderson et al. (2006).

As of 1 January 2020, planet Earth possessed only 63 localities where wild tigers were known to occur. These 63 Tiger Conservation Landscapes covered 911,920 km² across ten range states ( Figures 1 - 3 ,  4A , Table 2 ). They represent only 8.2% of the tiger’s likely resident indigenous range, suggesting a loss of approximately 91.8% of the world’s tiger habitat.

These 63 areas are the key landscapes for tiger conservation. By definition, each TCL had a high probability of resident tiger populations during the previous five years (i.e., between 1 January 2015 and 31 December 2019) and enough habitat with sufficiently low levels of human influence to sustain a population of at least five adult females. Designating an area as a TCL does not mean tigers exist everywhere within the mapped habitat block(s), however. Within the TCLs, only approximately a third of the total TCL area (311,698 km²) has been surveyed in the previous five years ( Figure 4B ).

Eight of the 63 TCLs identified in 2020 were transboundary. The largest TCL in the world was the Northern Triangle of India, Myanmar, China, Bhutan, and Bangladesh, which covered 294,847 km² in 2020, 32.3% of all TCL area worldwide. Unfortunately, tigers are unlikely to be continuously distributed throughout this vast area (Lwin et al., 2021; Sarkar et al., 2021; Sabu et al., 2022). Conflicts in Myanmar and India have constrained on-the-the ground surveys; the consequences of those conflicts for tigers and their prey remain poorly understood (Win, 2022).

All TCLs are important, but many were much smaller than the Northern Triangle. the Dawna Range, Ranthambore or Kaziranga TCLs, all cover less than 500 km². India had the largest number of TCLs in a single country at 35 landscapes due to fragmentation of what was once a much more continuous habitat but also concerted conservation efforts (Aylward et al., 2022). Nepal had the fewest TCLs (1), but a critical one, the inter-connected and extensive Terai Arc TCL of flooded grasslands and riparian forests, shared with India (Biswas et al., 2020; Thapa et al., 2021; Yadav et al., 2022). Other TCLs representing a diversity of habitats are found in the Russian Far East, parts of Southeast Asia, and on the island of Sumatra, as described below.

TCLs comprised only 50% of Asia’s effective potential habitat extant on January 1, 2020 ( Table 2 ). The remaining 50% was a combination of Restoration Landscapes, Survey Landscapes, or small fragments (tiger fragment, restoration fragment, or survey fragment; Table 1 ). Restoration Landscapes comprised 32% of the total effective potential habitat in 2020; Survey Landscapes, 8%, and all fragments, 9.3% ( Table 2 ). Fragmentation has been increasing over time, generating shards of habitat that are of little use to tigers but may provide future building blocks for connectivity.

Our classification of landscapes is based on 102,418 observations collected from 362 unique sources over the last 20 years ( Table S2 ). The majority of observations were of the “ad hoc” type, 55,684; sign survey observations, provided another 45,701 observations; and camera trap studies, 1,033 observations. By combining these observations with billions of satellite-based observations and human footprint calculations, we were able to follow range-wide trends in TCLs and related landscape types.

There are many ways to measure tiger habitat loss. The baseline deployed here is the “likely resident indigenous range”, an area of approximately 11.1 million km² where tigers are thought to have once lived (Sanderson et al., in press). This enormous range in Asia encompasses areas included in 30 modern countries, from Turkey to Russia, and China to Indonesia ( Table 2 ).

Since the time tigers were first significantly impacted by human beings, hundreds to thousands of years ago, up until 1 January 2020, tigers have lost an estimated 60% of the structural habitat, meaning that approximately 6.7 million km² of vegetative cover suitable for tigers has been lost through conversion to other uses such as agriculture or urban development. High levels of human influence have reduced the utility of another 2.5 million km² of structural habitat (23% of the indigenous range) through fragmentation, human-wildlife conflict, and likely concomitant reduction of the prey base. Direct persecution of tigers and/or tiger prey has further extirpated tigers from another 711,634 km² of suitable habitat (6% of indigenous range) that otherwise appears to be structurally sound and with low levels of human influence, as measured by the mapping of Restoration Landscapes. Regarding tigers, these areas can be considered “empty forests” (sensu Redford, 1992). Finally, lack of knowledge constrains our ability to understand tiger habitat, as measured by the Survey Landscapes. The status of tigers remains uncertain in 188,364 km² of habitat (2% of indigenous range), a decline of -75% since 2001. The decreasing amount of area needing survey reflects the increased effort to assess tiger status across Asia, but also underscores that there are fewer and fewer places to look.

Over the last 20 years, the total area of Tiger Conservation Landscapes (TCLs) declined from 1,025,488 km² in 2001 to 911,901 km² in 2020, a range-wide loss of 11% of TCL area ( Figure 4B ). TCLs also became more fragmented. The area of tiger fragments has increased 37%, from 3,477 km² in 2001 to 4,787 km² in 2020, though such fragments with tigers represent only a tiny percentage – less than 1% – of overall occupied habitat. Because of fragmentation, the number of TCLs has also increased, from 53 in 2001 to 63 in 2020.

These range-wide trends can be further decomposed at the national scale. In the following sections, we describe important differences in trends among tiger range countries over the first two decades of the 21^st century.

For tigers living in the wild, the news is not all bad. Encouragingly, this iconic species no longer teeters on the brink of extinction; an estimated 4,485 (3,726–5,578) tigers (Goodrich et al., 2022) live in the wild across ten countries in approximately a million square kilometers of habitat in the 63 Tiger Conservation Landscapes mapped here ( Table 2 ). Globally the loss of structural habitat has been largely arrested, albeit at historically low overall amounts (~40% of the likely resident indigenous range) ( Figure 4A ). Similarly, effective potential habitat now stands at only 16% of the indigenous range. Most encouragingly, metrics of occupied tiger habitat (as measured by TCL area) have declined only 11% in a two-decade period when Asia’s population grew 14% to 2.36 billion people, and the East Asian & Pacific economy grew 248% to 27.13 trillion US dollars (World Bank, 2023). Observed TCL areas fluctuated year-by-year ( Figure 4B ), largely as a function of the available knowledge base. Improved information may show less loss in the future. Taken together, our results suggest that tigers may have reached the bottom of the species recovery curve (Moorcroft, 2017). Tiger habitat is less tiger-filled and more fragmented than at any time in the last 200 years, but perhaps now the world is poised for a recovery in tiger populations not seen in generations.

This relatively optimistic reading of the trends at the global scale ( Figure 4 ) should not disguise the fact that tigers continue to lose habitat in many places, with different trajectories of loss in different nation-states ( Figure 5 ). The mechanisms of habitat loss at the grossest level are loss of vegetation (i.e. structural habitat) and expansion of human encroachment (i.e. reducing effective potential habitat). The expansion of linear infrastructure remains a concern for tiger conservation (Carter et al., 2020; Nayak et al., 2020). Concomitant declines in the prey base are probably also important but were not estimated here. (Modelling prey remains an important research topic.) During the last 20 years, tigers have become extinct in three Southeast Asian countries (Cambodia, Laos, and Vietnam) and across a large part of southeastern China. Tigers remain under pressure everywhere, which is especially acute in Southeast Asia. The tiger’s situation in vast parts of Myanmar remains unclear and is probably poor; ongoing conflict limits our ability to understand what is happening below the canopy. Human-wildlife conflict is frequent in Sumatra, leading often to harsh reprisals (Figel et al., 2021; Widodo et al., 2022; Patana et al., 2023). Malaysia’s tiger populations may have stabilized but at low levels, but how low remains unclear until the results of the recently finished country-level survey are released (Ding, 2022). Human influence appears much more harmful for tigers in Southeast Asia than in South Asia ( Figure S1 ). Though important questions about the situation remain, the evidence presented here suggests that without stepped-up conservation efforts, further losses are likely. Meanwhile, TCLs appear to be expanding in India, Nepal, Bhutan, northern China, and southeastern Russia.

Survey Landscapes have declined dramatically in the last 20 years ( Figure 4B ). Greatly expanded survey efforts, largely associated with increased use of remotely triggered camera traps (Karanth et al., 2004), have greatly increased scientific knowledge, allowing us to determine which areas with suitable habitat no longer have tigers (hence becoming Restoration Landscapes) or do still retain tigers, in which case they were incorporated as TCLs. The decline in area of Survey Landscapes represents our collectively improved knowledge of tiger distributions: we now know much more, even though there are fewer places where tigers persist. But because scientists often focus on repeated camera trap monitoring to detect trends in the best-known protected areas, large swaths of connected landscapes remain unsurveyed: typically, less than 50% of TCL area has confirmed sightings from the last five years. Therefore, even though we can confirm that a TCL retains tigers, in few cases, can we clearly understand what proportion of a TCL holds tigers and which areas are still in need of conservation efforts to allow populations to recover. These issues could use more attention in future iterations of the model.

TCL 3.0 demonstrates that there is still abundant effective potential habitat in Restoration Landscapes; these areas just lack tigers (Gray et al., 2023). If these extant habitats could be made suitable again in terms of prey base and were tigers to re-establish via natural dispersal or through active reintroduction efforts, we could increase the land base for tigers by 50%. Two conservation objectives – ensuring tigers are distributed at carrying capacity across the entirety of TCLs, and that they are recovered across Restoration Landscapes, could more than double, perhaps even triple, the number of free-living tigers on Earth (Lynam, 2010; Wikramanayake et al., 2011; Harihar et al., 2018). Given the tiger’s resilience and changes in Asia’s social and political landscape (Sanderson et al., 2019), this long-sought goal could be met within a human generation, if not sooner.

Models are only as useful as the data put into them. This model depends on in situ biological observations: surveys, with dates and locations, measures of effort, and verifiable evidence of tiger occurrence. The system doesn’t function without this information. Discovering a previously unknown or recently established tiger population is the easiest way to turn a landscape from green (Survey) to orange (TCL). On-going, extensive surveys in existing TCLs and Survey Landscapes are needed continuously to maintain the system and make its results useful.

Restoration Landscapes provide a remarkable, and until now, mostly unrealized opportunity to increase the area of occupied tiger range. Our analyses suggest that suitable habitat without tigers may be more susceptible to degradation and fragmentation than tiger-inhabited landscapes, as demonstrated by our analysis of what happened to former TCLs in Vietnam, Cambodia and Laos after extirpation. These findings highlight the urgency to halt deforestation, increase the prey base, and provide connection to TCLs to encourage dispersal. In some cases, where connectivity to an existent tiger population is not possible (e.g. the Balkash delta of Kazakhstan. some forest blocks in Thailand), active reintroduction efforts will be required. The Kazakhstan efforts seem particularly important because, were efforts to succeed, those tigers would be the first to live in Central Asia in more than a half-century.

Finally, there is – over some very long time scale – the issue of recovering some of the 60% of tiger indigenous range that has been lost, by redressing the conversion of former tiger structural habitat to other land use classes. Socioeconomic change in Asia, which has nothing per se to do with tigers, may provide enormous benefits to recovering tiger habitat (Sanderson et al., 2018; Sanderson et al., 2019). Rural-to-urban migration, urbanization, poverty reduction, and improved education may eventually make much more tiger habitat available, if (and only if) tigers remain in the wild and are capable of re-establishing in these areas. This is largely a civilizational project, to be enacted not only passively but proactively as national governments, civil society, and the global community pivot toward the realization that we are not separate from the Earth’s nature, but part of it, dependent as the tiger is, on the value of natural ecosystems for providing water, food and fiber, and a moderate climate. The United Nations Sustainable Development Goals and the recent “30 by 30” agreement adopted at the Convention on Biological Diversity CoP15 in Montreal (Einhorn, 2022), represent positive steps, which, if implemented well, can serve tigers, people, and the rest of nature.

Habitat maps and metrics are essential documents to plan species conservation efforts, reverse defaunation, and expand environmental protection amid the paired biodiversity and climate crises. National leaders must understand trends to make informed policy decisions and fulfil their international obligations ( Table 3 ). State and provincial officials need tabulations tailored to their local jurisdictions, with attention to protected areas, human-wildlife conflict, sustainable use, and connectivity to adjacent jurisdictions (see act-green.org). Non-governmental organizations and donors may be interested in trends of habitat availability range-wide ( Table 2 ) and in different ecologically defined areas (e.g. ecoregions, biomes), to inform ecological representation ( Table 4 ), or within collections of countries where they work ( Table 3 ). “On-the-ground”, landscape managers are most interested in the fine but pertinent details: Where is my landscape changing? How much habitat is available right now? Does the model match expectations or not? How does it relate to comparable areas?

Remarkably, existing technology can go a long way to serving all of these needs simultaneously. As described here, we have constructed not a map of tiger-relevant landscapes, but a system for mapping them. That system depends on clearly-defined inputs (an indigenous range map, species observations, a notion of what land cover are suitable for the species, and the human footprint) and delivers clear-defined outputs suitable for planning tiger conservation efforts at landscape, national, and range-wide scales. Because it is a mapping of the system, updates are possible by “simply” changing the inputs and re-running the process. Because it is a system when new inputs become available, for example, for 2021 or 2036, we can produce new, comparable maps and statistical outputs to measure progress.

One implication of a “mapping system” versus “a map” is that there is no one final map. Each new version is an improvement on the previous attempt, dependent on improvements in the accuracy and timeliness of the inputs, where within a few hours, a new observation can literally change the map for tigers. Versioning of results and linking back clearly to the inputs become even more important, even as we try to draw declarative statements about the status of tigers and trends in their habitat. This dynamic assessment system lies at the core of any adaptive management strategy.

A real constraint on improving understanding rests on the sharing of species observations. Conservationists and biologists have made huge gains in survey methods, repetition of tiger surveys, and documenting distribution in scientific papers and reports. But will people share the raw data they so laboriously collect? In this instance, despite a promising start with some commitments, enormous efforts and huge amounts of time were required to obtain the in situ biological observations required. After relentless emailing and Zoom calls failed to produce enough information, we resorted to a systematic literature search to re-capture and re-digitize information that sits on laptops across Asia and elsewhere. To develop a twenty-year analytical synthesis, we had to read every paper and report we could find published over the last 25 years and schedule dozens of follow-up conversations with individual researchers and research teams to cajole them into what should be the regular scholarly practice of sharing data for conservation, for example, using public Internet fora such as the Global Biodiversity Information Facility (Edwards, 2004; Ivanova and Shashkov, 2021) or Wildlife Insights (Ahumada et al., 2020). Many eventually complied and shared, including most of the co-authors of this paper (See also Table S1 ). One would expect scientific publishing to be one mechanism for better data sharing (Christensen et al., 2019). However, we found that the wide variety of ways that data are published in scientific papers and uneven review standards across journals ensure plenty of scope for researchers to dull the precision of the original data, hide or not report important details, and obfuscate the observations, even as authors claim to illuminate conservation issues. Not only are such abuses of the scientific literature process annoying, they ultimately delay decision-making, which is the ostensible reason for most studies in the first place.

Were such difficulties to be overcome, whether through ongoing, brute force literature searches or by better data sharing, then systems such as TCL 3.0 could help countries report on their commitments to the Sustainable Development Goal (SDG) agenda (United Nations Department of Economic and Social Affairs, 2022) and the Global Biodiversity Framework under the Convention for Biological Diversity (CBD Secretariat, 2022). Among other goals, SDG signatory countries have committed to Goal 15, Life on Land, which enjoins nations to protect, restore, and promote the sustainable use of terrestrial ecosystems, sustainably manage forests, combat desertification, and halt and reverse land degradation and halt biodiversity loss. TCL 3.0 enables countries to integrate ecosystem and biodiversity values into national and local planning processes, to understand how tiger conservation landscapes are protected and overlap with key biodiversity areas, and to halt forests loss since tigers are a species of forests in most parts of the range. Similarly, TCL 3.0 enables countries to measure the ecological integrity implied by and reinforced through long-lasting tiger populations.

Finally, we note the cultural aspects of technological innovations such as TCL 3.0 among conservationists. This project has depended on collaboration from across the tiger range, between governments and non-governmental organizations, and among the NGO sectors. Although tiger habitat has been much reduced, tigers live in too many places for any one actor or organization to successfully conserve them on its own. Rather technological tools such as this one, which depend on collaboration, also provide a framework to make the necessary collaborations desirable and rewarding. To this end, we have worked hard to link these results to other assessments, including the recently completed IUCN Red List assessment (Goodrich et al., 2022) and the ongoing IUCN Green Status assessment (Hunter et al., in prep.) for tigers. One important collateral product of TCL 3.0, which was not achieved through either of the previous versions, was a new range-wide vision for tigers written by and jointly endorsed by a coalition of conservation organizations (Coalition for Securing a Viable Future for the Tiger, 2022). The success metrics specifically prescribed under this shared vision ensure the continued application of the TCL 3.0 system into the future, which will reinforce the kinds of collaborations and mutual calls to action that tigers, and other wildlife, so richly deserve.