Soil scientists have traditionally regarded the deep, stable carbon reservoir stored in forest soils as a form of climate savings account — chemically locked, microbially unappealing, and preserved for centuries. A 37-year warming experiment conducted in the forests of central Massachusetts has just undermined that belief.
Following decades of consistent heating, the most resistant segment of soil organic matter at Harvard Forest started to decompose, emitting carbon dioxide that previous models had presumed would remain underground. The world’s longest-running soil warming study now indicates a feedback loop between microbes and climate that is more powerful, and slower to emerge, than what mainstream forecasts have indicated.
This discovery is significant as forest soils contain more carbon than the atmosphere and all living plants combined. If even a small portion of that reservoir becomes unstable under prolonged warming, the calculations of all emissions trajectories change.
A plot warmed for nearly forty years
The experiment commenced in 1991. Installed heating cables maintained a set of forest plots at Harvard Forest 5°C warmer than the surrounding soil, throughout the year, without any breaks. This temperature differential was selected to bracket the higher end of global warming estimates available at that time.
For perspective, global average temperatures have increased by approximately 1.1 to 1.4°C since the Industrial Revolution. The Harvard plots represent a glimpse into a future the Earth has yet to reach — but is moving towards.
Jerry Melillo has overseen the site since its inception. Over three decades, his team recorded the expected trend: microbes in the warmed soil operated more quickly, breaking down the easily digestible organic material and releasing CO2. That surge eventually diminished. Many researchers believed the system would stabilize into a new, warmer equilibrium.
It did not.
The surprise in the fourth decade
In the fourth decade of continuous warming, something unexpected occurred. The chemically stable carbon — the material tightly bound to minerals, enveloped in resilient molecular structures, and long believed to be off-limits to microbial consumption — started to decompose.
The peer-reviewed study in Science of the Total Environment captures the destabilization of persistent soil organic matter after about 30 years of warming. The enduring pool is not inert. It is merely waiting.
As explained by the Marine Biological Laboratory, Melillo noted that warming influences microbial communities in the soil, which play a crucial role in decomposing organic matter and can increase carbon loss. The research demonstrated that warming alters microbial communities in ways that can expedite the release of carbon from soils.
Why stable carbon was thought to remain secure
Soil carbon exists in various forms. Fresh leaf litter and root exudates are easy sources of energy — microbes consume them in weeks or months. Deeper down lies the recalcitrant fraction, sometimes hundreds or even thousands of years old, protected by physical barriers within soil aggregates and by chemical bonds to clay and iron minerals.
Climate models have usually represented this deep reservoir as a slow-turnover storage, adjusting mainly to land-use changes instead of just a few degrees of warming. The data from Harvard Forest challenges that perspective. Given enough time, warmer conditions seem to reshape microbial communities in ways that allow them to access carbon that was previously sequestered.
Though the mechanism is not entirely elucidated, the implicated pathway includes alterations in enzyme production, variations in the fungal-to-bacterial ratio, and the emergence of microbial specialists capable of breaking down mineral-associated organic matter. Essentially, the subterranean economy reconfigures itself.
The implications for climate calculations
Every significant climate model includes some version of a soil carbon feedback mechanism. Warmer soils release more CO2, which further heats the atmosphere, leading to warmer soils again. The ongoing question has always been how extensive the feedback is and how long it persists.
If short-term studies were utilized to calibrate that cycle — and most were, given that 37-year datasets are scarce — the models likely only captured the initial surge from labile carbon. The slower decay of stable carbon would have gone unnoticed.
Melillo and associates argue that integrating this delayed release into Earth system models will enhance forecasts of 21st-century warming. It will likely increase those projections as well.
The magnitude of the potential contribution is substantial. Global soils are estimated to store 1,500 to 2,400 gigatons of carbon in the top two meters, roughly two to three times the amount present in the atmosphere. A modest percentage loss over decades would mirror cumulative fossil fuel emissions from a large economy.
An earlier experiment, a familiar pattern
The result from Harvard Forest also contributes to an ongoing discussion about how ecosystems react to sustained CO2 and temperature pressure. A decade-long free-air CO2 enrichment study in the Mojave Desert, published in Nature Climate Change, revealed that arid ecosystems can accumulate significantly more carbon under elevated CO2 than previously acknowledged. That study identified deserts as unrecognized carbon sinks.
When considered together, the two experiments illustrate a more intricate balance. Some ecosystems accumulate more carbon than anticipated when CO2 levels rise. Others, once sufficiently warm for a prolonged duration, begin to deplete reserves thought to be untouchable.