Beneath the surface of a California conifer forest, an extraordinary experiment has been quietly running for more than seven years. Resistance cables buried in the soil have held entire soil profiles at a steady four degrees Celsius above ambient temperature, from the litter layer down to a meter deep. Now, a team of researchers has taken an unprecedented look at what that sustained warmth is doing to the microscopic communities that control how carbon moves through the ground, and their findings are reshaping what scientists thought they knew about the planet’s largest terrestrial carbon reservoir.
The study, published in Nature Communications, combined deep metagenomic sequencing with metaproteomics, an approach the authors call proteogenomics, to reconstruct more than 2,300 microbial genomes from soil samples collected across the full 0 to 90 centimeter depth range. The work took place at the University of California’s Blodgett Forest Research Station in the Sierra Nevada foothills, where three pairs of heated and unheated plots have been monitored since 2014. Over 7.5 years, the team collected 115 soil samples at six time points, generating 3.61 terabases of raw sequencing data and detecting more than 25,000 proteins expressed in situ. The resulting dataset spans 48 classes across 23 phyla, including many lineages that have never been cultured in a laboratory.
The single most striking result is that soil depth, not warming, is the dominant force shaping microbial life underground. Depth explained roughly 33 percent of the variance in community composition, while four degrees of warming accounted for just 2.3 percent, a fourteenfold difference. The sharpest ecological transition occurred between 30 and 50 centimeters, a zone where soil organic matter chemistry, root inputs, and redox conditions all change abruptly. This finding matters because more than half of the global soil carbon stock lies below 30 centimeters, in subsoils that have long been neglected by both field studies and the biogeochemical models used to project climate futures.
The trait analysis revealed a microbial world organized by resource gradients. Surface communities were enriched in genes for methanol, formaldehyde, and formate oxidation, reflecting their access to plant-derived methanol and pectin compounds from root exudates and litter. Methanol dehydrogenases appeared in 620 genomes, representing 27 percent of the reconstructed genomes across 13 phyla. In contrast, deep-soil microbes showed greater potential for carbon monoxide oxidation, alcohol fermentation, nitrogen fixation, and hydrogen metabolism, strategies consistent with life where oxygen and favorable substrates are scarce. Proteomic measurements confirmed that these patterns are not merely genomic potential: carbon monoxide oxidation and nitrogen fixation proteins peaked below 50 centimeters, while methanol oxidation was active at every depth sampled.
Perhaps the most unexpected discovery involved sialidases, enzymes that cleave terminal sialic acids from glycoproteins. These GH33-family enzymes, along with their associated carbohydrate-binding modules, were among the most abundant carbohydrate-active enzymes detected in deep soils, and the team believes this is the first evidence of their expression in subsoil environments. Sialic acids are known to protect glycoproteins from proteolytic degradation and thermal denaturation, meaning that microbial necromass rich in these sugar-coated proteins has often been considered a relatively stable carbon pool. The abundant expression of sialidases suggests that deep-soil bacteria actively strip away this protective layer, unlocking carbon and nitrogen from fungal and microbial remains that were previously thought to be locked away.
Warming effects, though smaller than the depth signal, were functionally consequential and strikingly depth-specific. After 7.5 years, surface soils showed enhanced potential for transporting amino sugars and sugar alcohols, mobilizing phosphorus, and detoxifying reactive oxygen species, alongside an expanded repertoire of carbohydrate-active enzymes targeting cellulose, xylan, and sialic acids. In mid-depth soils between 10 and 60 centimeters, warming stimulated formaldehyde detoxification, formate oxidation, and carbon monoxide oxidation. Nitrogen cycling shifted in complex, bidirectional ways: nitric oxide reduction potential increased near the surface, nitrate and nitrous oxide reduction rose at mid-depth, ammonia oxidation potential deepened in the profile, and dissimilatory nitrate reduction to ammonium declined.
These functional shifts appear to be driven by carbon limitation rather than changes in bulk soil chemistry. Previous measurements at the site showed that warming reduced fine root mass and depleted unprotected, easily accessible soil organic carbon in the subsoil. Phospholipid fatty acid analysis revealed that microbial biomass below 30 centimeters declined by 28 percent under warming, even as bulk carbon measurements showed no change. The researchers argue that deprived of plant-derived substrates, deep-soil microbes are turning to alternative metabolisms, including carbon monoxide oxidation and single-carbon compound turnover, to sustain themselves. This functional plasticity, rather than taxonomic turnover, appears to be the primary mode of microbial adaptation to long-term warming.
The study also highlighted the outsized roles played by poorly understood, uncultured lineages. Dormibacteria, Gemmatimonadota, and Thermoplasmatota contributed hundreds of near-complete genomes and were disproportionately associated with single-carbon metabolism and nitrogen cycling, particularly in deep soils where proteomic data confirmed their activity. These organisms are largely invisible to standard amplicon-based surveys that rely on primer-dependent amplification, underscoring the value of genome-resolved approaches for capturing the true functional scope of soil ecosystems. Their prominence in deep soils suggests that some of the most important players in carbon-climate feedbacks have been hiding below the reach of conventional sampling.
The findings arrive at a critical moment for climate modeling. Most ecosystem biogeochemistry models neither explicitly represent depth-dependent heterogeneity nor incorporate microbial traits, a gap that contributes to large uncertainties in projected carbon-climate feedbacks. The Blodgett experiment has already documented a sustained 30 to 37 percent increase in soil respiration under warming, including proportional responses from subsoils, and this new mechanistic framework helps explain why. Surface microbiomes respond rapidly through substrate uptake and stress tolerance, while deep communities respond more slowly but with functionally distinct strategies that may sustain greenhouse gas fluxes over longer timescales. The authors argue that the next generation of Earth system models must move beyond topsoil-centric perspectives and build in the depth-stratified microbial traits that this study has now mapped in unprecedented detail.
What emerges from seven and a half years of heated ground is a portrait of soil as a vertically stratified ecosystem, where each layer harbors microbes with fundamentally different survival strategies. As the planet warms, the carbon locked in deep soils will not simply decay faster in a uniform way; instead, distinct communities at each depth will shift their metabolisms along different trajectories, some scavenging trace gases, others dismantling protective glycoproteins, and still others rerouting nitrogen through removal pathways. Understanding these layered responses, the researchers conclude, is essential for predicting whether the vast carbon reserves beneath our feet will remain stable or become an accelerating source of atmospheric carbon dioxide in a warming world.
Subject of Research: Depth-stratified microbial carbon cycling responses to whole-soil warming in a temperate forest soil profile
Article Title: Soil proteogenomics uncover microbial strategies for soil depth-specific carbon cycling under whole-soil warming conditions
Article References: Soil proteogenomics uncover microbial strategies for soil depth-specific carbon cycling under whole-soil warming conditions. (n.d.). https://doi.org/10.1038/s41467-026-78137-0
Image Credits: AI Generated
DOI: 10.1038/s41467-026-78137-0
Keywords: soil carbon, whole-soil warming, proteogenomics, metagenomics, metaproteomics, microbial traits, subsoil, carbon cycle, nitrogen cycling, sialidases, climate feedbacks, soil microbiology
Cite Scienmag News
Violet Maxwell. (October 13, 2026). Deep Soil Microbes Reveal Hidden Carbon Strategies Under Seven Years of Warming. Scienmag. https://scienmag.com/deep-soil-microbes-reveal-hidden-carbon-strategies-under-seven-years-of-warming/
Violet Maxwell. "Deep Soil Microbes Reveal Hidden Carbon Strategies Under Seven Years of Warming." Scienmag, 13 October 2026, https://scienmag.com/deep-soil-microbes-reveal-hidden-carbon-strategies-under-seven-years-of-warming/. Accessed 13 October 2026.
Violet Maxwell. "Deep Soil Microbes Reveal Hidden Carbon Strategies Under Seven Years of Warming." Scienmag. October 13, 2026. https://scienmag.com/deep-soil-microbes-reveal-hidden-carbon-strategies-under-seven-years-of-warming/

