Warming and Nitrogen Pollution Are Draining a Hidden Carbon Reservoir Beneath Drylands

Beneath the grasslands and shrublands that ring the world’s arid and semi-arid regions lies a carbon reservoir that climate models have largely treated as inert: soil inorganic carbon. Locked into minerals such as calcium carbonate, this vast pool rivals the better-known organic carbon stored in plant matter and humus, and in many dryland soils it actually exceeds it. Now an eight-year field experiment in a semi-arid grassland on China’s Loess Plateau, paired with a meta-analysis of nitrogen-addition studies, shows that this overlooked reservoir is far more fragile than assumed. When climate warming and nitrogen pollution arrive together, the mineral carbon bank begins to bleed.

The study, published in Nature Geoscience by Tongshuo Bai of Nanjing Agricultural University and colleagues, examined how the three most pervasive global change drivers—warming, reactive nitrogen input and altered precipitation—reshape the two carbon pools simultaneously. The team manipulated these factors in open-top chambers and rainfall shelters across the growing seasons from 2016 to 2023, tracking soil inorganic carbon, soil organic carbon, nitrogen transformations and soil acidity through the full profile of the top 40 centimeters. The result was a clear and unsettling asymmetry: precipitation changes alone left the inorganic pool untouched, nitrogen addition alone left it unchanged, but the combination of warming and nitrogen triggered a measurable decline of roughly 9.1 percent in soil inorganic carbon.

That interaction matters because it exposes a mechanism hidden in single-factor experiments. In the warmed plots that also received nitrogen, the researchers documented enhanced microbial nitrogen mineralization and nitrification—the two-step process by which microbes convert ammonium into nitrate. Nitrification releases hydrogen ions, acidifying the soil. In calcareous soils, that acidity is neutralized by dissolving carbonate minerals, a reaction that pulls calcium carbonate apart and releases carbon as bicarbonate or carbon dioxide. In other words, the very microbial machinery that nitrogen fertilization and warming rev up produces the protons that dissolve the mineral carbon stock. The loss was not an accident of weathering but a direct, predictable consequence of nitrogen-driven acidification accelerated by heat.

The trade-off revealed by the experiment is the study’s most striking finding. Nitrogen input and warming, the two drivers with the strongest effects, pushed soil inorganic carbon and soil organic carbon in contrasting directions. As the mineral pool declined, the organic pool shifted in the opposite manner, producing a see-saw relationship between the two reservoirs over the eight years of observation. Meta-analysis of 44 field observations from sites across China confirmed the pattern, showing that nitrogen addition and warming induce contrasting changes in the two carbon pools beyond the Loess Plateau site. This means that bookkeeping which credits nitrogen deposition with boosting organic carbon sequestration in grasslands may be silently subtracting an equivalent or larger debit from the inorganic ledger.

The scale of the potential error is enormous. Recent global assessments estimate that soil inorganic carbon amounts to well over a trillion tonnes worldwide, concentrated in the drylands that cover roughly 40 percent of the land surface. A separate synthesis found significant losses of inorganic carbon at the continental scale in China, and topsoil carbonate stocks across China’s grasslands declined substantially between the 1980s and the 2000s. Yet bibliometric analyses show that inorganic carbon is systematically ignored in global soil carbon research, which has historically equated soil carbon with organic carbon. Earth system models that project land carbon sinks under warming scenarios rarely represent carbonate dissolution at all, treating it as a geological process too slow to matter on human timescales.

The new evidence suggests that assumption is wrong. Carbonate dissolution can respond to environmental change within years, not millennia. Warming accelerates microbial activity, which accelerates proton production, which accelerates dissolution. Previous work has shown that nitrogen fertilization raises carbon dioxide efflux from inorganic carbon in croplands worldwide, and that nitrification inhibitors can alleviate the dissolution of soil inorganic carbon caused by nitrogen fertilization. The Loess Plateau experiment extends this picture to a natural grassland ecosystem and demonstrates the interactive amplification: neither warming nor nitrogen alone moved the pool, but their combination produced losses consistent with the acidification pathway. Drought may further exacerbate dryland inorganic carbon loss under warming, adding a third axis of vulnerability that the current study did not observe directly but that the broader literature increasingly supports.

The chemistry explains why the pool is so sensitive to its microbial surroundings. Calcium carbonate dissolves when the soil solution becomes less alkaline, and the protons generated by nitrification are consumed directly in that reaction. The researchers found that the inorganic carbon loss co-occurred with enhanced exchangeable calcium dynamics and shifts in the fraction of calcium-bound and water-soluble carbon, linking the mineral pool to the organic pool through calcium-mediated stabilization. Calcium acts as a glue for organic matter, bridging minerals and organic compounds into protected aggregates. When acidification strips carbonate from the soil, it does not just release inorganic carbon; it also undermines the calcium scaffolding that helps stabilize organic carbon, potentially destabilizing both reservoirs at once. The see-saw observed in the field data may therefore reflect coupled biogeochemistry rather than two independent processes moving in opposite directions.

For dryland management, the implications cut in uncomfortable directions. Reactive nitrogen now enters ecosystems from fertilizer application, livestock operations and fossil fuel combustion, and deposition rates are rising across much of the world’s dry interior. Efforts to increase grassland productivity through nitrogen amendment are often justified by the organic carbon they lock into soils. The new results caution that such gains may be partially or wholly offset by inorganic carbon losses, particularly in calcareous soils where carbonate is abundant and vulnerable. Conversely, the identification of nitrification as the proton source points toward mitigation: practices that slow nitrification, such as inhibitor use or management that reduces nitrogen inputs, could protect the mineral carbon stock. Organic amendments that promote new carbonate formation offer another route to rebuilding it.

The experiment’s eight-year span also reveals the temporal texture of the change. Soil inorganic carbon content tracked against its 2015 baseline showed the divergence emerging as treatments accumulated their effects, and depth-resolved sampling in 2023 and 2024 captured the responses across the upper profile. Total soil carbon—the sum that most monitoring programs measure—masked the underlying redistribution, which is precisely why the trade-off has gone unnoticed. A soil can appear to be gaining or holding carbon while quietly swapping one pool for another, with very different residence times, vulnerabilities and climate feedbacks attached to each.

What the study ultimately delivers is a warning about hidden liabilities in the carbon budget. Drylands are projected to expand and warm, nitrogen deposition will continue to climb, and the inorganic carbon pool—large, long-lived and previously assumed safe—now appears to be a live participant in the Anthropocene carbon cycle. Incorporating carbonate dynamics into models of dryland carbon storage will require field measurements like those from the Loess Plateau, along with the recognition that soil acidity, microbial nitrogen cycling and mineral dissolution form a connected chain that global change is pulling taut. The other carbon, as its advocates have called it, can no longer stay out of the ledger.

Subject of Research: Responses of soil inorganic and organic carbon to warming, nitrogen input and altered precipitation in dryland ecosystems

Article Title: Trade-offs between soil inorganic and organic carbon driven by multiple global environmental changes

Article References: Bai, T., Wang, Y., Deng, J., Qiu, Y., Zhang, H., Hall, S. J., Zhang, Y., & Hu, S. (2026). Trade-offs between soil inorganic and organic carbon driven by multiple global environmental changes. Nature Geoscience. https://doi.org/10.1038/s41561-026-02135-6

Image Credits: AI Generated

DOI: 10.1038/s41561-026-02135-6

Keywords: soil inorganic carbon, soil organic carbon, drylands, climate warming, nitrogen deposition, nitrification, soil acidification, carbon cycle, grassland ecosystems, carbonate dissolution, Nature Geoscience, carbon sequestration