Researchers have produced the first global map of underground fungal networks, revealing a living web of threadlike structures stretching roughly 110 quadrillion kilometers through Earth’s topsoils. That distance is more than a billion times the span between Earth and the sun. The mapping effort, built from 322 studies and more than 16,000 soil cores, estimates these networks weigh about 300 megatons and channel roughly 4 billion tons of carbon dioxide equivalent into soils each year, a volume equal to about 11 percent of annual human-related CO2 emissions. The findings carry direct consequences for how scientists and policymakers think about carbon storage, agricultural soil health, and the hidden cost of intensive farming.
Why a 110-quadrillion-kilometer fungal map changes the carbon math
The scale of these networks reframes a basic question: how much carbon is the ground already absorbing without human intervention? The study, published in the journal Science, focused on arbuscular mycorrhizal (AM) fungi, organisms that form partnerships with plant roots and extend hair-thin filaments called hyphae through topsoil. Those filaments pull carbon from the atmosphere and lock it into the soil matrix. The researchers calculated that the annual CO2 equivalent flow through these networks sits at roughly 4 billion tons, a figure that makes fungal infrastructure one of the largest biological carbon pumps on the planet.
That 11 percent share of human emissions is not a rounding error. It means any large-scale loss of fungal networks directly weakens a natural carbon sink that operates at industrial scale. The data show that croplands already carry roughly 50 percent lower AM fungal densities compared to less disturbed soils. Grasslands, by contrast, hold about 40 percent of the total global fungal network length, according to the Society for the Protection of Underground Networks. The gap between those two land types points to a measurable cost of intensive agriculture on below-ground carbon cycling.
One hypothesis worth tracking is that fungal network density could show a measurable recovery threshold once pesticide application rates drop below about 1.5 kilograms per hectare in previously intensive cropland zones. That threshold has not been tested in the current dataset, but the 50 percent density gap between croplands and less disturbed soils suggests a steep response curve. Repeated soil-core sampling over three growing seasons in transitioning fields could reveal whether reduced chemical inputs trigger detectable regrowth, or whether degradation has already passed a point of easy reversal.
How 16,000 soil cores and machine learning built the map
The research team compiled field measurements from 322 separate studies, drawing on more than 16,000 soil cores sampled from the top 15 centimeters of soil across multiple continents. From that dataset, they used machine-learning models to predict AM fungal density and biomass in areas where direct sampling had not occurred. The result was a global picture of network length, estimated at approximately 1.10 × 1017 kilometers, and a biomass figure of roughly 300 ± 60 megatons.
Earlier work published in the journal Nature had begun linking fungal density to carbon flux, but the Science paper extended that foundation into a spatially explicit global map. The machine-learning approach allowed the team to move beyond scattered field observations and generate predictions for regions with sparse sampling histories. That methodological step is what turned a collection of local measurements into a planet-scale estimate, though it also introduces modeling uncertainty that the biomass error margin of plus or minus 60 megatons partially reflects.
The 40 percent concentration of fungal infrastructure in grasslands carries practical weight for land-use planning. Grasslands are often considered lower-value real estate compared to forests in carbon accounting frameworks. This data suggests that converting grassland to cropland does not just release above-ground carbon. It also strips out a dense underground network that was actively pulling CO2 into the soil. For farmers and land managers, the implication is concrete: soil that has lost half its fungal density is soil that stores less carbon and likely supports less efficient nutrient cycling for crops.
Those same maps also highlight how unevenly fungal networks are distributed. Regions with intact grasslands and low-intensity grazing emerge as hotspots of underground biomass, while heavily tilled, chemically intensive agricultural belts show marked declines. Because AM fungi trade nutrients for plant-derived sugars, any disruption to plant diversity, root structure, or soil chemistry can reverberate through the hyphal network. The new map does not yet provide farm-by-farm prescriptions, but it offers a first-order guide to where conservation or restoration of fungal infrastructure could yield the largest climate and soil-health benefits.
What the fungal map cannot yet answer
Several gaps remain in the evidence. The 300-megaton biomass estimate lacks independent ground-truth measurements from outside the study’s own dataset. The 322 contributing studies vary in methodology, geography, and sampling depth, and the published results aggregate those inputs into global totals without breaking out regional accuracy or confidence intervals by continent. That means the map is strongest as a global average and weakest as a guide to any specific farm, watershed, or national park.
The 4-billion-ton CO2 equivalent flux figure comes from downstream summaries of the research rather than from a fully public methodological breakdown. The connection between fungal density and actual carbon storage rates involves assumptions about hyphal turnover, soil chemistry, and plant–fungus exchange rates that are still being refined. Independent replication of the flux estimate, using different modeling approaches or direct isotope-tracing field experiments, has not yet appeared in the published record.
The recovery hypothesis also remains untested. While the 50 percent density gap between croplands and less disturbed soils suggests that reducing disturbance and chemical inputs could allow networks to rebound, the time scales are unknown. Hyphal strands can regrow quickly under favorable conditions, but long-term degradation of soil structure, organic matter, and microbial diversity may limit how far and how fast recovery can proceed. Without longitudinal field trials that track both management changes and fungal metrics, it is impossible to say whether a simple reduction in pesticides or tillage will restore the original carbon sink capacity.
Another open question is how climate change will interact with these underground systems. Rising temperatures, shifting rainfall patterns, and more frequent droughts will alter plant communities and soil moisture regimes, both of which shape AM fungal activity. The current map effectively freezes the world at a single point in time. To inform climate policy, researchers will need dynamic models that couple fungal networks with changing vegetation, land use, and atmospheric CO2 levels, capturing feedbacks rather than static snapshots.
For now, the map’s most robust message is qualitative: Earth’s topsoils host an immense, previously undercounted infrastructure for moving carbon out of the atmosphere and into the ground. That infrastructure is sensitive to how humans farm, graze, and develop land. Protecting and rebuilding these fungal networks will not replace the need to cut fossil fuel emissions, but it could strengthen one of the planet’s quietest and most extensive natural carbon sinks. The challenge for scientists is to refine the numbers; the challenge for policymakers is to decide how much this hidden web should shape future land-use and climate strategies.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
