SPUN Maps Earth's Hidden Fungal Network: 110 Quadrillion Kilometres of Living Threads

• A landmark study published in Science has produced the first global estimate of arbuscular mycorrhizal fungal networks, finding roughly 110 quadrillion kilometres (about 68 quadrillion miles) of living fungal threads in the top 15 centimetres of Earth's soils.

• Researchers from SPUN,  Vrije Universiteit Amsterdam, and more than a dozen institutions combined data from over 16,000 soil samples with robotic imaging of 300,000 individual fungal strands to build machine-learning maps at one-kilometre resolution.

• The findings reveal that grasslands, not tropical rainforests, hold the densest fungal networks on Earth, and that croplands are associated with roughly 47% lower fungal density, a pattern with significant implications for soil carbon storage and agricultural resilience.

The Invisible Circulatory System Beneath Our Feet

Beneath every meadow, prairie, and forest floor is a vast and vital infrastructure almost no one knows exists. Arbuscular mycorrhizal (AM) fungi, microscopic organisms that form partnerships with around 70% of all land plant species, weave dense networks of thread-like cells called hyphae through the soil. These hyphae, each 10 to 50 times thinner than a human hair, shuttle phosphorus and nitrogen to plant roots in exchange for carbon from the plant. According to the study, plants allocate an estimated one billion tonnes of carbon per year to AM fungi, making this exchange a significant driver of the global carbon cycle.

Until now, no one had reliably measured how much of this fungal infrastructure actually exists. A new paper in Science, authored by Justin D. Stewart, Corentin Bisot, E. Toby Kiers and colleagues across institutions including SPUN, Vrije Universiteit Amsterdam, the University of Sheffield, AMOLF, Oxford University, and Hasselt University, has changed that.

Credits: SPUN

Building the First Global Map

The research team assembled a database from 322 studies representing more than 16,000 soil cores across 100 ecoregions worldwide. They trained Random Forest machine-learning models on this dataset, correlating environmental variables such as soil chemistry, vegetation cover, and climate with measured fungal densities. The models predicted AM hyphal density at approximately one-kilometre resolution across all vegetated land on Earth.

To convert those density estimates into biomass, the team used a custom-built imaging robot at the AMOLF Institute in Amsterdam that captured more than 300,000 width measurements across five strains of three species — Rhizophagus clarus, Rhizophagus irregularis, and Rhizophagus sp. isolate 165. The average hyphal radius came out at 2.7 micrometres, enabling the team to calculate volume and, from there, weight.

Their estimate: approximately 110 quadrillion kilometres of living AM hyphae in the top 15 centimetres of soil, with a total biomass of around 300 million tonnes of carbon which is roughly four to six times the biomass of all humans alive today.

Surprisingly, Grasslands- Not Rainforests- Dominate Below Ground

Perhaps the most striking finding concerns where these fungi are densest. Despite tropical forests being among the world's most productive ecosystems above ground, the study found that grasslands contain the highest predicted AM hyphal densities, approximately 39% denser than tropical moist forests. Grassland ecosystems, including the Tibetan Plateau, the Everglades in Florida, and the Sudd wetlands in South Sudan, together account for around 40% of predicted global AM hyphal biomass. The authors suggest this reflects the tendency of grasses to allocate more carbon to fungal partners than woody plants do.

Croplands tell a less encouraging story. Observed field data showed that agricultural soils had, on average, almost half the AM hyphal densities than non-cropland soils, a pattern the authors link to elevated phosphorus and nitrogen inputs and fungicide application, both of which can suppress fungal growth.

Credits: SPUN

Why This Matters for Climate Initiatives

The headline numbers are striking, but the climate implications are the real story here. AM fungi are an active carbon sink. Through their exchange with plants, these fungi draw down the equivalent of roughly 4 billion tons of CO2 annually, about 11% of fossil fuel emissions. That's a flux on par with major national emissions totals, happening quietly underground every year.

This is also why the cropland finding matters so much. It's not just "farms have less fungi" — it's that these networks function as living infrastructure that draws carbon into soils and supports much of life on Earth, and intensive agriculture appears to be eroding that infrastructure at scale. A ~50% reduction in hyphal density across the world's croplands isn't a local soil-health footnote; given how much land is under cultivation globally, it represents a meaningful chunk of potential carbon-sequestration capacity that's been suppressed.

The grassland finding cuts the same way. Ecosystems like the Tibetan Plateau, the Everglades, and the Sudd wetlands are places not typically centered in climate-mitigation conversations the way tropical forests are, and may be doing outsized work belowground. If grasslands are conserved or restored with carbon in mind, this data suggests fungal networks (not just plant biomass) should be part of that accounting.

The authors are careful to flag what's still unknown: turnover rates- how fast this fungal carbon is cycled versus locked away, remain poorly constrained, which matters enormously for translating biomass into actual sequestration estimates. Biomass tells you how much carbon is tied up in fungal tissue right now; it doesn't tell you how long it stays there, which is the number climate modelers actually need.

As SPUN's Toby Kiers put it, fungi have largely been left out of climate and conservation planning, and this dataset is pitched as a first step toward changing that, by giving conservation planners and carbon-policy people an actual map to work from rather than guesswork.

Source: Stewart et al. (2026), "Global density and biomass of arbuscular mycorrhizal fungal networks," Science, 392(6803), pp. 1171–1176. https://doi.org/10.1126/science.adu4373