As global temperatures rise, they could trigger repeated surges of carbon dioxide emissions from forest soils, and, in a worst-case scenario, create runaway global warming, a long-term experiment carried out in a New England research forest shows.

The unexpected pulses of CO2 emissions are coming from evolving communities of microorganisms, including fungi and bacteria, that break down carbon. Scientists have known that these microbes become more active in warmer temperatures, and they have theorized that CO2 emissions would spike with rising temperatures and then decline as the microbes deplete carbon in the soil. But the repeated surges of CO2 emissions over time from several artificially heated research plots came as a surprise.

The experiment, beneath oaks and maples in the Harvard Forest in Massachusetts, started in 1991, when scientists heated 18 test plots of forest ground to 5 degrees Celsius above the temperature of adjacent areas and started to measure how the soil changed.

In a study published this week in the journal Science, the scientists describe what they’ve learned so far and discuss the connections between the warmer soil and the surprising up-and-down cycle of CO2 emissions, which holds clues to changes ahead in warming future. 

They said they nearly ended their observations when CO2 losses tapered off after 10 years, but then a second pulse of emissions started. The second surge suggests that the microbes in the soil are evolving to be able to break down more resistant mineral- and wood-based sources of carbon, the scientists said.

There have been signs of similar responses in other regions, raising concerns about a “long-term self-reinforcing carbon feedback from mid-latitude forests” that could pump more CO2 into the atmosphere and intensify global warming, the scientists wrote.

Exactly how much CO2 could be released and how fast it would be pumped out is still the subject of intense research, and the results from the controlled experiment at Harvard Forest are difficult to extrapolate to a wide scale. But the findings fit similar patterns documented by different types of studies at forests in Sweden and from long-term ecological monitoring in southern Germany.

The Microbes are Evolving

Similar microbial changes in other landscapes would signal a persistent climate feedback loop, said the study’s lead author, Jerry Melillo, a climate scientist at the Woods Hole Marine Biological Laboratory. More warmth activates more microbes that can process tough sources of carbon, leading to more emissions and more warming, he explained.

Co-author Kristen DeAngelis, a microbiologist at the University of Massachusetts-Amherst, said new species of microbes are appearing. She also traced an enzyme fingerprint suggesting that the microbes are evolving to adapt to new, poorer and less abundant carbon in the soil.

“There is a chance of this mechanism feeding itself,” Melillo said. “The microbes are ubiquitous, and if they respond in the same general pattern, we may not be able to switch them off.”

Soils are the largest repository of organic carbon in the terrestrial biosphere, containing an estimated 3,500 billion metric tons of carbon. By comparison, human activities release about 36 billion metric tons of carbon annually into the atmosphere. Estimates of global CO2 emissions from soil vary, but a 2016 study estimated that human-caused warming would account for 12 to 17 percent of the expected emissions from soils through 2050.

The Harvard Forest study found that, since 1991, the heated plots have lost 17 percent of the carbon that had been stored in the top 60 centimeters of soil. The research showed that, for the first 10 years, micro-organisms in the soil—thousands of species of bacteria and fungi—went into overdrive, breaking down carbon and releasing it to the air as heat-trapping CO2. Then, carbon loss from the heated plots quickly dropped to the same level as from the unheated control area for eight years, before surprising scientists by climbing to another five-year spike from 2008 to 2013, followed by yet another decline.

“The point of this study is that it’s very long term. The carbon losses we measured are not based on extrapolation from short-term lab experiments,” Melillo said. After measuring the first up-and-down cycle, he said the experiment was almost shut down. “We wrote that we thought we were exhausting the available carbon and maybe entering a period of no accelerated emissions,” he said, referring to an earlier study.

New Pulses Change Global Warming Equation

The details of the forest soil carbon cycle are important on a global level for figuring out how much and how fast Earth will continue to warm, said Markus Reichstein, who was not involved in the study but does similar research as director of the Max Planck Institute for Biogeochemistry.

Most global warming calculations take into account a first pulse of CO2, like that documented at the Harvard Forest test plots, but not a second and potential future increases, which could really change the equation, he said.

“What the researchers here seem to have been able to do is to detect changes in microbial community profiles linked to the depletion of different soil carbon pools,” he said. That will help predict on a larger scale the complex “cascading interactions between changes in vegetation and other organisms.”

This is important because many early climate change projections suggested the biosphere would sequester more carbon, said University of Stirling biogeochemist Jens-Arne Subke, who was not involved in the study.

“Recent research like the Harvard Forest study are starting to suggest that warming forests could start to emit more carbon than they absorb,” he said. It’s unclear exactly when that tipping point could be reached, but more detailed and long-term studies will help answer the question, he said.