Some of those prehistoric droughts in the northern Great Plains of what is now the United States also lasted longer than modern-day dry spells such as the 1930's Dust Bowl decade, according to sediment core studies by the team.
The group's evidence implies these ancient droughts persisted for up to several decades each. At their heights, prairie fires became uncommon because there was too little vegetation left to burn. The ages of charcoal deposits suggest instead that prairie fires occurred during intervening wet periods, with each wet-dry cycle lasting more than a century each.
A report on the research will be delivered at a session at 1:30 p.m. on Wednesday, Aug. 4, in Meeting Room D136 of the Oregon Convention Center during the Ecological Society of America's 2004 annual meeting in Portland.
"We were looking for the effects of past climate changes on ecosystems," said Jim Clark, H.L. Blomquist Professor of Biology at Duke's Nicholas School of the Environment and Earth Sciences. But when Clark and his colleagues began examining evidence from the mid-Holocene period of 5,000 to 8,000 years ago in parts of the Dakotas, Montana and western Minnesota, "nothing seemed to make any sense."
"The question was: Could we look at the sediments for charcoal evidence of the amount of fire, for pollen evidence of the kinds of grasses that were growing then, for sediment chemistry to show how much erosion was going on, and be able to deduce climate changes -- or the lack of them -- under way at the time?" Clark said.
When he and his colleagues finally determined the correct intervals between samplings was about once every decade, "the patterns just jumped right out at us," he recalled in an interview. "We were seeing these very coherent drought cycles.
"What would happen was that the grass would disappear. So the fuel for fire would be lost. We'd see the erosion start. The chemistry of the lakes would change. We would see these dust-bowl effects.
"And then, within several decades to a century later, the grasses would come back, fires would start back up and erosion would stop."
To make these deductions, Duke post-doctoral investigator Kendrick Brown evaluated prehistoric charcoal deposits. Joe Donovan, a geophysicist at the University of West Virginia, studied the geochemistry of the soil samples. Eric Grimm and Pietra Mueller of the Illinois State Museum in Springfield investigated pollen in the sediments.
The regularity of these ancient droughts make much more recent Great Plains droughts in the 1890s and 1930s appear "unremarkable" by comparison, Clark said, even though the contemporary ones "walloped people."
The study did not speculate how the findings might relate to anticipated future climate change, when a surge of carbon dioxide from human activities is predicted to cause Earth's climate to warm appreciably.
"What we can say that is relevant is that these sort of drought cycles are common and most of the climate models predict increased aridity in continental interiors in the future," Clark said.
"One could speculate that the droughts could be all that much worse when you realize that it's not only climate change from changing CO2 content in the atmosphere, but also this natural variability out there that we don't fully understand."