When compared to grassy areas, Jackson's research group found the extra carbon saved in the wood of encroaching trees and shrubs at the wettest locations may be more than offset by the carbon lost from the underlying soil.
They made their deductions at six experimental field sites from eastern Texas to western New Mexico, where original species of grasses and invading species of woody vegetation could be directly compared.
"Assessments relying on carbon stored from woody plant invasions to balance emissions may therefore be incorrect," Jackson and four co-authors wrote in the Aug 8, 2002 edition of the journal Nature.
Jackson, an associate professor at Duke's Biology department and Nicholas School of the Environment and Earth Sciences, referred to an article in the June 22, 2001 issue of Science.
That Science report highlighted "woody encroachment" as one of two important factors for its estimate that the atmospheric carbon "sink" - amounts of former carbon dioxide gas stored away in various natural repositories - had been "relatively stable" in the United States for more than a decade.
Jackson and his colleagues set out to examine the carbon sink associated with woody plant encroachment in more detail.
Many scientists consider carbon dioxide from human activities a major contributor to global warming. Policy makers have hoped that growing more trees would help ameliorate the problem.
That's because some of the carbon from the carbon dioxide is incorporated into the plant material, where it remains until the plants die and decay. In the case of the wood in trees, that carbon may thus remain sequestered for centuries. In the case of grasses, carbon from the plant matter will return to the atmosphere in only a matter of years.
However, another sink for carbon is the soil. The "organic carbon" in soils "can stay there for a long time, longer than in the plant," Jackson said in an interview. "It can remain in the soil for centuries. Furthermore, the global soil carbon pool is about twice as large as the plant pool."
So while grassy vegetation in itself is no long-term carbon repository, the rich black soils underneath many grasslands can be. And it's the grasslands once carpeting the Southwestern U.S. that are changing as a result of fire suppression and cattle grazing, he said.
Many of those former grassland environments have since been invaded by drought-tolerant woody tree and shrub species that grow more robustly in the eastern part of their range and more sparsely towards the drier west.
So Jackson's research group -- also including Jay Banner of the University of Texas at Austin, Jackson's graduate student Esteban Jobbagy, William Pockman of the University of New Mexico, and Diana Wall of Colorado State University - focused on he soil underlying what had been southern and western grasslands.
"This kind of analysis for grassland soil carbon was something new," Jackson said. "It had been proposed that the woody species might even increase soil carbon compared to the grasslands. People really didn't think that grasslands would store more carbon in the soil than woodlands."
In research supported by the National Science Foundation and Andrew W. Mellon Foundation, Jackson's group first examined global records comparing amounts of carbon in grasslands, shrublands and woodlands in various climates and environments worldwide.
That data search found that "as you move to increasingly wet environments, grasslands have a lot more soil carbon than shrublands and woodlands do," he said. "That was somewhat of a surprise. The analysis suggested that sites with the potential to store the most plant carbon also had the potential to lose the most soil organic carbon."
The scientists then set up experiments at six research sites in Texas, New Mexico and Colorado, where grasslands had been invaded by trees and shrubs over the past century.
They were able to study the effects of this encroachment because the original grasses had been maintained on one side of a fence running through each test plot, while woody species had invaded the other side of the fence.
The average precipitation levels at those plots spanned "the whole range of grasslands in North America," he said, from the edge of the western deserts to the edge of the eastern forests.
The group also probed soils more deeply than previous studies have with the aid of a drilling rig than could penetrate as much as 10 meters down. And they were able to deduce the original sources of the soil carbon because carbon taken in by woody vegetation is different from that processed by grasses.
"We found a clear negative relationship," the authors wrote in Nature, between the amount of precipitation and changes in soil organic carbon "when grasslands were invaded by woody vegetation." Drier sites consistently gained soil organic carbon, while wetter sites lost it.
Loses of "organic" soil carbon at the wetter sites were "substantial enough" to offset the increased "plant biomass" carbon stored in the growing wood, they reported..
Why this precipitation-tied carbon loss is occurring is still unclear, Jackson said "We don't know the exact mechanism yet, but we have some suggestions." One is that "grasslands send a lot of their carbon below ground, so that carbon goes immediately into the soil," he noted.
In addition to changes in the amount of carbon entering the soil, the quality of the tissue also changes," he added. "Woody tissue is typically more difficult to decompose than herbaceous tissue."
The Nature report said "the most notable shift in carbon from below-ground to above-ground pools was at the wettest site," in Engeling, Tex. "Such shifts make carbon stocks more vulnerable to loss from fire, biomass (wood) harvesting and other disturbances."