India’s citizen science movement needs backup

Come April, armed with phones binoculars, people across India will turn their gaze to nature. From roadside trees to city lakes, thousands join the City Nature Challenge, a four-day global bioblitz to record as many species sightings as possible.

Since 2018, the Indian edition of the Challenge (CNC) — which accounted for nearly one-third of global participation in 2024 — has generated thousands of observations, uploaded to iNaturalist, the world’s largest biodiversity-recording platform.

Observations that reach ‘research grade’ are automatically shared weekly with the Global Biodiversity Information Facility (GBIF), making them accessible for scientific analysis, conservation planning and policy.

The surge in participation has yielded notable finds — from the first report of stiletto flies (Megapalla species) in India to a sea slug rediscovered after 163 years.

However, the limited availability of experts to review and validate submissions has hampered the conversion of these observations into research-grade, scientifically reliable data, even as under-sampled countries like India stand to gain most from integrated crowdsourced data, studies say.

“We have a yawning gap between the volume of biodiversity data and the capacity to pin it down to research-quality standards,” says Ram Dayal Vaishnav, head of Citizen Science and Engagement at The Naturalist School, Bengaluru.

Vaishnav and colleagues tracked CNC observations across 12 cities between 2023 and 2025¹. The team followed observations through iNaturalist’s quality-control pipeline to see where they stalled. They found 121,988 observations stuck in the ‘Needs ID’ category, revealing what they describe as a mounting curation crisis among a small group of specialists working beyond sustainable capacity.

Roughly 20% languish at the genus level and cannot be confidently assigned to species. In 2024 alone, observations grew by 286%, while the pool of expert identifiers shrank by 13%.

Verification success also varies sharply across taxa. Birds, supported by strong and active expert communities, saw more than 80% of records reach research grade. Fungi, plants and insects, by contrast, accumulate in the ‘Needs ID’ queue — photographed and geotagged, but awaiting confirmation.

Citizen science for understudied regions

Despite being one of the most biodiverse countries, India has contributed only about 240,000 reliably georeferenced, non-cultivated vascular plant records to GBIF — less than 0.1% of the global total of 338.9 million. Chronic underinvestment in science has played a role.

However, since 2015, India’s GBIF contributions have grown by roughly 35% annually, largely driven by iNaturalist, which now supplies more than half of the country’s records, notes Daniel Lusk of the University of Freiburg in Germany.

By combining 300 million citizen science observations — including 200,000 from India — with expert vegetation surveys, global trait databases, and satellite data on climate, soil and vegetation, Lusk and colleagues built AI-powered Global Trait Maps² charting 31 key plant traits at one-square-kilometre resolution.

Researchers can now analyse how properties such as leaf size and plant growth height vary globally, helping measure climate resilience. "The unique environmental and vegetative qualities of the Indian subcontinent and its more than 200,000 citizen science observations certainly aided in producing more robust models with lower extrapolation error," says Lusk.

Different data sources contribute complementary strengths. Structured surveys capture grasses and understory plants often missed by photography-based platforms, while citizen science excels at documenting trees and flowering species, reflecting a bias toward conspicuous, easily photographed species in accessible areas.

Together, they expand both species coverage and geographic reach. But with around 14 GBIF records per plant species, compared with hundreds or thousands in Europe, India's trait maps rely more on model extrapolation and less on direct observations.

Trait models, however, are only as reliable as their validation datasets. “Crowdsourced, expert-coordinated vegetation surveys remain the best available benchmark,” Lusk notes.

Thomas Vattakaven, programme manager at the India Biodiversity Portal, a tech-driven platform that has been bringing together biodiversity information since 2008 , believes crowdsourced species interaction data, largely overlooked in current projects, could propel ecology into an entirely new realm of understanding.

“Trait-based insights, such as those from the Global Trait Maps group, are a wellspring of research ideas. The story doesn’t end with simply ‘finding’ them.”

That potential is evident in the case of the tree-climbing Malabar tree toad. A modelling study³ harnessing citizen science data from the Mapping Malabar Tree Toad programme under the Portal’s Frogwatch — which helps assess the conservation status of many lesser-known amphibian species — showed that the species is not as rare as previously thought, though neither is it common. “It may have been underreported rather than truly absent. But it is climate-sensitive, patchily protected, and facing habitat redistribution,” says Vattakaven.

Other platforms, such as the 15-year-old SeasonWatch, are shifting toward what rigorously collected datasets can predict, and who might ultimately benefit.

The project looks at the effects of seasons on tree phenology — participants register neighbourhood trees and observe leaves, flowers, and fruit weekly. A phenology dataset representing a substantial proportion of the project’s database, uploaded to GBIF, has been cited about 97 times in the two years since its release, says project head Geetha Ramaswami.

Other data from the platform have attracted remote sensing scientists seeking ground-based validation for satellite-derived patterns, while ecological analyses have uncovered statistically significant links between tree phenology and environmental variables, opening pathways toward predictive tools for farmers.

"If you're a mango grower, you might want to know how rainfall and temperature patterns this year could affect flowering or fruiting," explains Ramaswami.

Bottlenecks holding back India’s biodiversity data

To truly scale citizen science–driven research and conservation, critical institutional, knowledge and infrastructure gaps must be closed, experts say.

Vaishnav argues for cultivating both ‘broad identifiers’, who can refine observations to narrower taxonomic groups, and ‘last-mile’ experts who can confirm species-level identifications. Local workshops connecting observers with specialists could accelerate validation.

Investing in the digitisation of biodiversity archives — museum collections and specimen databases — would create robust training datasets for accurate AI-based species identification, similar to systems in Europe, North America and Australia, notes seed ecologist Shyam Phartyal of Mizoram University in Aizawl, India, and a contributor to the plant trait maps.

Phartyal, an avid citizen scientist, sees opportunity in engaging experienced taxonomists — retired botanists, forest officers and senior scientists — who have deep expertise but may hesitate to adopt digital tools. “We need to find ways to bring them into the system,” he says.

Training contributors is equally critical. Ramaswami emphasises explaining why protocols matter. Mycologist Sanjay Singh at Agharkar Research Institute recommends designing training materials to assist enthusiasts in better photographing fungi with a focus on anatomical features. High-resolution photos showing diagnostic details — flower parts, wing venation, gill structure — are often key to successful identification, adds Vaishnav.

Sustained funding is another necessity. Maintaining digital portals, ensuring data security and improving user interface require stable backing. “We need the support of the IT sector,” says Vattakaven.

At its core, the challenge of ecological data lies in separating signal from noise. Natural systems rarely conform to tidy rules, and citizen science introduces biases in sampling and coverage, Lusk adds.

When these platforms were first conceived, the emphasis was largely on democratising science by bringing people into the data-collection process. Privacy, safety, attribution and data quality are now central concerns. The IUCN’s formal recognition of citizen science at its 2025 global meeting reflects this shift, alongside growing emphasis on FAIR principles — ensuring data are Findable, Accessible, Interoperable and Reusable, so they can integrate into infrastructures such as GBIF and inform policy.

“Today, it would not be acceptable to start a citizen science programme without thinking about these issues right at the outset,” concludes Ramaswami.