Global synthesis of 423 studies reveals biodiversity’s uneven grip on nature’s services

The most comprehensive synthesis of its kind ever assembled has delivered a verdict that will reshape how scientists and policymakers think about the living world: biodiversity underpins the functioning of ecosystems and the services they provide to humanity, but it does so with startling unevenness. An international team led by researchers at Imperial College London and King’s College London compiled 423 studies containing more than 222,000 individual data points, drawn from terrestrial, freshwater, marine and estuarine systems across every inhabited continent. Published in Nature Ecology & Evolution, the analysis is more than twice the size of the next largest database in the field, and it moves the debate beyond the simple question of whether biodiversity matters to the far more consequential questions of where, how and under what circumstances it matters most.

The scale of the undertaking reflects the urgency of the problem it addresses. Human pressures such as land-use change and the spread of exotic species are estimated to have pushed around one million species towards extinction, threatening the pollination, carbon storage, water purification and food production systems on which societies depend. Global frameworks, including the Kunming-Montreal Global Biodiversity Framework, aim to halt these declines, but effective action demands a quantitative understanding of exactly how diversity translates into ecological performance. The new study follows the classification scheme of the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services, distinguishing ecosystem functions, the processes occurring within ecosystems, from ecosystem services, the contributions those processes make to human well-being, while treating carbon sequestration separately across land and sea because of its central role in climate regulation.

One of the study’s central findings concerns the shape of the relationship between biodiversity and ecosystem performance. Ecologists have long debated whether adding species yields steadily increasing benefits, a linear relationship implying that every species, including rare ones, contributes meaningfully, or whether benefits plateau quickly, a saturating relationship suggesting widespread functional redundancy. Across 23 categories of functions and services, the team found that quadratic models best described 11 categories and linear models 9, with only two best fit by logarithmic curves. In other words, linear and saturating forms are almost equally common, and for many services the benefits of diversity continue to climb across the biodiversity gradients observed in the empirical literature. The authors argue that saturation limits have rarely been reached in the studies compiled, implying that functional redundancy may be systematically overestimated and that individual species contribute more uniquely than commonly assumed.

Where saturation does occur, the thresholds differ dramatically between realms. In freshwater systems, service delivery typically saturates at a median of roughly four species, with a median absolute deviation of two, whereas terrestrial systems require around 14 species and marine systems around 23. The researchers attribute this gradient to fundamental differences in ecological architecture: freshwater food webs tend to be dominated by generalists operating at relatively small scales, while marine systems, exemplified by coral reefs, combine vast spatial extent, high trophic complexity and pervasive mutualisms and co-evolutionary interactions that demand a larger species pool to sustain function. Marine ecosystems may also display low species multifunctionality, meaning individual services can be supported by a small subset of species, but maintaining the full portfolio of services requires high overall diversity.

To compare effect strengths across these different functional forms, the team calculated standardized effect sizes, using Fisher’s Z-transformed correlation coefficients for each of 1,959 datasets. The results were emphatic: higher biodiversity enhanced most functions and services, from carbon capture to food production, but the magnitude varied enormously. Regulation of air quality showed an effect size of essentially zero, while oceanic carbon sequestration registered a value of 1.48, by far the strongest response of any category examined. This exceptional sensitivity of oceanic blue carbon, a significant component of the global carbon cycle, suggests that marine carbon cycling, including phytoplankton productivity, microbial loop dynamics and the biological carbon pump, is under tight biological control linked to community composition and trophic structure. The authors caution, however, that oceanic sequestration is represented by only seven datasets in the database, compared with 154 for terrestrial sequestration, a critical knowledge gap that itself constitutes a call to action.

Not every service behaves the same way. Hazard regulation, the capacity of ecosystems to buffer floods, storms and erosion, showed weak overall sensitivity to biodiversity despite often displaying linear relationships in the functional-form analysis. The explanation lies in functional uniqueness: some services are delivered overwhelmingly by a few foundational species or ecosystem engineers. Sand dunes, for instance, are stabilized by one or two key shrub species that create a natural barrier against flooding and erosion, and adding further species does not necessarily improve that protective performance. Yet the wider community still matters, because the persistence of these foundational taxa depends on the ecological conditions that a diverse community maintains. The study connects this pattern to the concept of ecosystem-specific planetary boundaries, thresholds beyond which ecosystem functions can collapse, and proposes developing a typology of species and system traits to help conservationists identify in advance which systems are governed by uniqueness and which by redundancy.

Context proved decisive throughout the analysis. Using linear mixed-effects models, the team found that biodiversity effects generally strengthened with increasing spatial grain, consistent with beta-diversity effects in which spatial turnover in species composition enhances functional complementarity across sites. But the pattern reversed for specific services: contributions to hazard regulation, net primary productivity and biomass turnover all declined significantly at coarser scales, echoing forest studies where positive relationships at fine grains turn negative at broader ones. Ecosystem type also mattered, with biodiversity effects positive on average in freshwater systems, negative in marine systems and non-significant in terrestrial ones, though these averages concealed strong variation among service categories. Observational studies revealed modestly but statistically stronger effects than experiments, suggesting that small-scale, simplified experimental plots may underestimate the cumulative, context-dependent impacts of biodiversity loss that play out across real landscapes.

The synthesis then ventures beyond description into forecasting. Because the underlying literature is heavily biased towards developed countries, underrepresenting regions critical for global food production such as South America and Africa, the team built a heuristic modelling framework linking their biodiversity-ecosystem service relationships with projected biodiversity intactness under the IPCC’s Shared Socioeconomic Pathways. Applied to agricultural pest regulation, measured as the ratio of natural enemies to pests, a widely used proxy for top-down biological control, the model projects that pest regulation declines more severely under the fossil-fuel-driven SSP5 scenario than under the middle-of-the-road SSP2, with reductions forecast across North America, South America and Asia. Crucially, the declines are disproportionately concentrated in countries experiencing rapid population growth and low Human Development Index values, precisely the nations most reliant on local ecosystem services and least equipped to substitute pesticides or imports for lost natural pest control.

The implications reach well beyond agriculture. By quantifying where biodiversity loss will bite hardest, the framework gives planners a scientifically grounded basis for anticipating problems before they materialize, and the team has released model outputs and forecasts for all 23 service and function categories across five socioeconomic scenarios. The authors are candid about limitations: the projections assume that species loss is random with respect to functional importance, whereas if vulnerable species are also functionally critical, future service declines could be steeper than predicted. Progress, they argue, will require hybrid networks embedding experimental manipulations within long-term monitoring, hierarchical sampling to separate local, regional and landscape diversity contributions, and a new generation of field experiments designed explicitly to disentangle redundancy from uniqueness. What the synthesis establishes beyond reasonable doubt is that biodiversity is not a luxury of pristine wilderness but the operating infrastructure of a functioning planet, and that the richest communities, those closest to saturation thresholds never yet observed, are the ones humanity can least afford to lose.

Subject of Research: Global relationships between biodiversity and ecosystem functioning and services

Article Title: Biodiversity safeguards ecosystem services and functions worldwide

Article References: Moffett, E. R., Gayford, J. H., Chen, L., Morris, O. F., Shi, Y., Somekh, L., Stasik, N., Purvis, A., Woodward, G., & Pearse, W. D. (2026). Biodiversity safeguards ecosystem services and functions worldwide. Nature Ecology & Evolution. https://doi.org/10.1038/s41559-026-03200-4

Image Credits: AI Generated

DOI: 10.1038/s41559-026-03200-4

Keywords: biodiversity, ecosystem services, ecosystem functioning, carbon sequestration, blue carbon, pollination, pest regulation, IPBES, shared socioeconomic pathways, functional redundancy, marine ecosystems, global synthesis