Grasslands hold the densest fungal networks on Earth, not rainforests.
Beneath meadows and prairies, a living infrastructure older than civilization has been quietly cycling carbon on a planetary scale — and science has only now found a way to see it. Researchers from SPUN, Vrije Universiteit Amsterdam, and partner institutions have produced the first global map of arbuscular mycorrhizal fungal networks, revealing 110 quadrillion kilometres of hyphal threads sequestering the equivalent of roughly 4 billion tonnes of CO2 each year. The map overturns a long-held assumption: it is grasslands, not rainforests, that harbour the densest fungal life on Earth, while the world's croplands have lost nearly half their fungal density to industrial agriculture. Humanity has been making decisions about land, carbon, and climate without knowing this circulatory system existed at scale — and that ignorance now has a shape.
- The urgency is quiet but immense: a carbon sink rivalling the emissions of entire nations has been operating invisibly underground, unaccounted for in climate models or conservation policy.
- The disruption cuts against decades of assumption — tropical rainforests have long anchored the imagination of carbon conservation, but grasslands like the Tibetan Plateau and the Sudd wetlands are now revealed as the true strongholds of fungal biomass.
- Industrial agriculture is suppressing this infrastructure at planetary scale, with cropland fungal density running nearly 47 percent below that of non-agricultural soils — a loss tied directly to fertilizers and fungicides.
- Scientists are navigating the gap between discovery and utility: the map exists, but fungal turnover rates remain unknown, leaving climate modelers unable to convert biomass into reliable long-term sequestration projections.
- The study lands as a foundation rather than a solution — a first coordinate in a territory that conservation planners and policymakers have never before had the tools to navigate.
Beneath every meadow and prairie lies a living infrastructure almost no one has measured. Arbuscular mycorrhizal fungi — microscopic partners to roughly 70 percent of all land plants — spin thread-like hyphae through soil ten to fifty times thinner than a human hair, trading phosphorus and nitrogen for carbon in a transaction so fundamental that plants send an estimated one billion tonnes of carbon annually into these underground partnerships.
A team from SPUN, Vrije Universiteit Amsterdam, and more than a dozen other institutions has now produced the first global accounting, published in Science. Drawing on over 16,000 soil samples across 100 ecoregions and machine-learning models trained on soil chemistry, vegetation, and climate data, the researchers mapped fungal density at one-kilometre resolution across all vegetated land. A custom imaging robot at Amsterdam's AMOLF Institute measured more than 300,000 individual fungal strands to convert predictions into biomass. The result: approximately 110 quadrillion kilometres of living fungal threads in the top 15 centimetres of soil — around 300 million tonnes of carbon, four to six times the biomass of every human alive.
The most striking finding overturns conventional wisdom. Grasslands, not tropical rainforests, hold the densest fungal networks on Earth — roughly 39 percent denser than tropical moist forests. Ecosystems like the Tibetan Plateau, the Everglades, and South Sudan's Sudd wetlands account for around 40 percent of all predicted global fungal biomass, likely because grasses allocate more carbon to fungal partners than woody plants do — a pattern largely invisible to climate and conservation planning until now.
The agricultural picture is sobering. Croplands show fungal densities nearly 47 percent lower than non-agricultural soils, a pattern linked to fertilizer inputs and fungicide use. At planetary scale, this represents a substantial suppression of living carbon-sequestration infrastructure. The fungi collectively draw down the equivalent of roughly 4 billion tonnes of CO2 annually — about 11 percent of global fossil fuel emissions — happening silently underground every year.
Yet critical unknowns remain. Fungal turnover rates — how long carbon stays locked in soil rather than cycling back — are poorly constrained, and that is precisely the figure climate modelers need for long-term projections. The study is framed as a first step: giving planners an actual map rather than educated guesses. Whether it will be enough to shift how the world values grasslands, farmland, and the invisible systems beneath our feet remains an open question.
Beneath the surface of every meadow and prairie lies a vast living infrastructure that almost no one has ever seen or measured. Arbuscular mycorrhizal fungi—microscopic organisms that partner with roughly 70 percent of all land plants—spin thread-like networks through soil so fine that individual filaments are ten to fifty times thinner than a human hair. These fungal strands, called hyphae, ferry phosphorus and nitrogen to plant roots in exchange for carbon, a transaction so fundamental that plants send an estimated one billion tonnes of carbon annually into these underground partnerships. Until recently, no one had any reliable sense of how much of this fungal infrastructure actually existed on Earth.
A team of researchers from SPUN, Vrije Universiteit Amsterdam, and more than a dozen other institutions has now produced the first global accounting. Published in Science, the study assembled data from over 16,000 soil samples collected across 100 ecoregions and used machine-learning models trained on environmental variables—soil chemistry, vegetation, climate—to predict fungal density at one-kilometre resolution across all vegetated land. To convert those predictions into actual biomass, the team deployed a custom-built imaging robot at Amsterdam's AMOLF Institute that measured the width of more than 300,000 individual fungal strands from three species. The result: approximately 110 quadrillion kilometres of living fungal threads in the top 15 centimetres of soil worldwide, representing roughly 300 million tonnes of carbon—four to six times the total biomass of every human being alive.
The most surprising finding upends conventional thinking about where Earth's most critical carbon-cycling infrastructure lives. Grasslands, not tropical rainforests, contain the densest fungal networks on the planet—roughly 39 percent denser than tropical moist forests. Ecosystems like the Tibetan Plateau, the Everglades, and the Sudd wetlands in South Sudan together account for around 40 percent of all predicted global fungal biomass. The researchers suggest this reflects the way grasses allocate more carbon to their fungal partners than woody plants do, a pattern that has gone largely unnoticed in climate and conservation planning.
The agricultural picture is grimmer. Croplands show fungal densities roughly 47 percent lower than non-agricultural soils—a pattern researchers link to elevated fertilizer inputs and fungicide application, both of which suppress fungal growth. This is not merely a local soil-health concern. Given the scale of global cultivation, a near-50 percent reduction in fungal density across the world's farmland represents a substantial loss of potential carbon-sequestration capacity. The fungi are an active carbon sink, drawing down the equivalent of roughly 4 billion tonnes of CO2 annually—about 11 percent of global fossil fuel emissions, a flux that rivals the total emissions of major nations, happening silently underground every year.
Yet the researchers are careful about what remains unknown. Turnover rates—how quickly fungal carbon cycles through the system versus how long it stays locked in soil—remain poorly constrained. Biomass tells you how much carbon is currently tied up in fungal tissue; it does not tell you how long that carbon remains sequestered, which is the figure climate modelers actually need to project long-term carbon storage. This gap between what the map reveals and what policy makers require to act on it is significant. The study is framed as a first step: giving conservation planners and climate policy makers an actual map to work from rather than educated guesses. Whether that map will be enough to shift how the world thinks about grasslands, farmland, and the invisible circulatory systems beneath our feet remains to be seen.
Citas Notables
Fungi have largely been left out of climate and conservation planning, and this dataset is pitched as a first step toward changing that.— Toby Kiers, SPUN
La Conversación del Hearth Otra perspectiva de la historia
Why does it matter that grasslands have denser fungal networks than rainforests? Aren't rainforests still the priority for climate work?
Because we've been looking at the wrong place. Rainforests get attention because they're visibly productive—tall trees, obvious biodiversity. But the carbon work happening underground in grasslands is actually denser. If we're trying to protect carbon-sequestration infrastructure, we've been neglecting the place where it's most concentrated.
And the cropland finding—47 percent less fungal density—what's actually causing that? Is it the fertilizer, the fungicides, or something else?
Probably both. When you add synthetic nitrogen and phosphorus, plants don't need to invest as much carbon in their fungal partners to get the nutrients they need. Fungicides directly kill the fungi. Together, they've essentially dismantled the living infrastructure that was doing the carbon work for free.
Four billion tonnes of CO2 annually—that's a real number, right? Not speculative?
It's based on the carbon flux through the fungal-plant exchange, so it's grounded in measured biology. But here's the catch: that's the carbon moving through the system right now. We don't know how much of it actually stays in the soil versus gets respired back into the air. That's the turnover-rate problem. You can have a massive carbon flow and still not have permanent sequestration.
So the map is useful but incomplete for climate policy?
Exactly. It's the first honest picture of where this infrastructure exists and how much of it we've already damaged. But translating that into actual climate projections requires understanding persistence, and we're not there yet.