A sweeping new analysis of global biodiversity data has found that when a species starts disappearing from the places it once reliably turned up, that local retreat is often an early signal of a much bigger problem: a heightened risk of extinction.

The findings arrive against a backdrop of well-documented global biodiversity loss. A landmark United Nations assessment in 2019 estimated that roughly one million species face extinction in the coming decades, the fastest rate of biodiversity decline in human history. Conservation scientists have long needed faster, cheaper ways to flag which species deserve closer attention, since formal extinction risk assessments are resource-intensive and inevitably lag behind on-the-ground change.

The new study, published today in Nature Communications by researchers at the University of St Andrews and an international team of collaborators, draws on one of the most extensive long-term ecological datasets ever assembled. Using BioTIME, a biodiversity database developed at St Andrews, the researchers tracked more than 60,000 populations belonging to 2,362 species across 978 marine and terrestrial assemblages, each monitored for at least 20 years.

A wildlife assemblage is a community of species from the same broad taxonomic group — fish, birds, insects and so on — that share a particular place and time. By comparing how often individual populations within these assemblages appeared or disappeared over the decades, the team could measure something subtler than raw abundance: how consistently a species was actually present where ecologists expected to find it, a metric known as local prevalence.

That measure turned out to carry a meaningful signal. The researchers cross-referenced each population’s prevalence trend with its species’ official conservation status on the International Union for Conservation of Nature’s Red List, the global standard for extinction risk. Populations that were becoming locally scarcer over time were disproportionately likely to belong to species already classified as more vulnerable to extinction.

The pattern, however, was far from uniform. The team also identified plenty of confounding cases: populations of officially threatened species that had remained stable or even grown locally, alongside populations of supposedly secure species that were quietly declining. Overall, fewer than 10 percent of the populations studied showed a clear long-term increase or decrease in prevalence at all — most fluctuated without a strong directional trend.

“Our results highlight the importance of assemblage-level monitoring in conservation strategies and show the value of long-term time series such as those in BioTIME,” said Dr. Faye Moyes of the University of St Andrews’ School of Biology, one of the study’s two joint lead authors.

Her co-lead author, Laura Antão of the University of Turku in Finland, said the project marked the first time researchers had tested for a consistent link between population-level prevalence trends and extinction risk status using broad assemblage-monitoring data, rather than relying solely on species-by-species assessments. “Finding a clear signal that decreasing prevalence is usually associated with a higher risk of extinction is a strong indication that we can detect impacts even for species that are not currently classified as at risk,” Antão said.

That distinction matters because formal Red List assessments don’t cover every species on the planet — far from it. Assemblage monitoring data, much of it collected for general ecological research rather than conservation triage, already exists in large quantities and could potentially flag species sliding toward trouble well before a formal risk assessment catches up.

Senior author Anne Magurran, a professor at St Andrews, framed the findings as a possible early-warning tool. “These temporal trends could serve as early-warning indicators and help target both new monitoring efforts and conservation actions,” she said. “For instance, stable populations of at-risk species are of key conservation interest, while declining trends might highlight species that are missing from extinction risk assessments.”

The researchers stress that the relationship between local prevalence and global extinction risk is genuinely complex, shaped by factors that vary across species, ecosystems and geography. A population can decline locally for reasons that have nothing to do with a species’ overall survival prospects, and a species’ global status can mask wide variation in how individual populations are faring on the ground. But the overall trend — decreasing prevalence tracking with elevated risk — held up across the dataset’s enormous geographic and taxonomic range, from coral reef fish to grassland birds to forest insects.

Study co-author Maria Dornelas, also a professor at St Andrews, noted that the sheer scale of the task facing biodiversity science makes combining existing datasets especially valuable. “Because the task of assessing biodiversity change is gigantic, and we cannot travel in time to collect more data in the past, we want to use all information available,” Dornelas said. “These two large biodiversity databases have only limited overlap, and this study shows us how we can leverage both to expand our understanding of biodiversity change.”

The findings arrive as global environmental change continues to reorganize ecological communities across nearly every biome, pushing some species toward range contractions and population collapse while others adapt or even expand. By showing that long-term ecological monitoring data — collected for entirely different scientific purposes — can double as an extinction risk detector, the St Andrews-led team has opened a potential shortcut for conservation planners working with limited time and resources.

Whether that signal proves reliable enough to directly inform conservation triage decisions will likely depend on follow-up research testing it across additional ecosystems and longer time horizons. For now, the study adds weight to a growing case that the steady, unglamorous work of long-term ecological monitoring — counting the same fish, birds and insects in the same places, year after year — may be one of conservation science’s most underused early-warning systems.

Endnotes

Journal: Nature Communications

Article title: “Linking species local trends from assemblage monitoring to global extinction risk”

DOI: 10.1038/s41467-026-74132-7

Source: University of St Andrews, via EurekAlert!, June 23, 2026