Forward into the past

The 1970s brought the United States a string of fierce winters and a spate of speculation on a cooling climate. Many atmospheric scientists had a different worry: they knew that carbon dioxide in the air had been increasing for decades and that global temperatures should rise before long. New physical evidence, streaming in from fossils, polar ice, and other proxies, all pointed toward sharp climatic swings in the past. Motivated by these clues, a diverse set of scientists—several based at NCAR, with others using the center’s resources—worked on paleoclimate modeling that would test the evidence.

After spending a semester teaching at the Lamont-Doherty Earth Observatory in 1976, Stephen Schneider returned to NCAR hungry to expand the scope of climate research along interdisciplinary lines. He later spoke about “the absolute essential need to use the paleoclimatic record as the backdrop against which we calibrate our understanding of the tools that we use to make predictions of the future.”

Photo of a group of men standing in front of a supercomputer
In applying global models to study Earth's past climate, Eric Barron (third form left) worked with leading developers of the NCAR Community Climate Model, including (left to right) Warren Washington, Maurice Blackmon, Bob Chervin (background), and Byron Boville.

At the same time, Eric Barron was spending summers in Boulder on NCAR fellowships while working on his doctorate in oceanography from the University of Miami. For his dissertation, Barron chose a formidable task: determining whether a global model could simulate the climate of the mid-Cretaceous period, around 100 million years ago, when dinosaurs and ferns flourished near the Arctic Circle. Barron ended up teaming with Schneider and colleague Starley Thompson to see whether the location of continents could explain the Cretaceous warmth. In an influential 1981 paper in Science, the three researchers announced that the era’s warmth was consistent with as much as seven times the modern-day concentration of carbon dioxide. However, the possible sources of that CO2 weren’t clear, and the model struggled to keep polar regions as warm as the evidence indicated they had been. “It is much easier to explain a Cretaceous with warmer tropics and cooler poles,” they wrote.

Other paleoclimate modelers set their sights on more recent periods. At the University of Wisconsin–Madison, Bette Otto-Bliesner—having built a low-resolution global model for her dissertation there—teamed up with UW’s John Kutzbach, a regular NCAR visitor. They examined whether the so-called Milankovitch cycles (regular variations in Earth’s orbit around the Sun) could be enough to trigger climate change. The two suspected that about 10,000 years ago, when Earth was closest to the Sun in July, the intensified summer sunlight at northern latitudes would have strengthened the African and Asian monsoons. If so, that would help explain evidence that lakes and grasslands had once covered the Sahara.

“These data were becoming available, but nobody had tested them. That was our opportunity, and we seized it,” recalls Kutzbach. The initial results he obtained with
Otto-Bliesner prompted them to use NCAR’s fledgling Community Climate Model for a more in-depth study. After the first CCM runs showed a dramatic enhancement of the monsoons, the team expanded on those results in a years-long series of studies. “As NCAR developed its models, we could dig deeper and deeper,” Kutzbach says.

Several other modelers in the paleoclimate community spent formative years at NCAR. To examine past climates with richer detail, Starley Thompson and David Pollard built a pioneering offshoot of the CCM that incorporated vegetation and other land-surface features. Thompson later moved to Lawrence Livermore National Laboratory, while Pollard joined Barron in an paleoclimate group at Pennsylvania State University.

Barron himself—one of the first to use computer models to peer into the atmosphere’s past—didn’t completely foresee what his own future would bring. In 2008, he was hired to direct NCAR, and in 2010 he became the 14th president of Florida State University.


Today — Life, death, and the ice age

Photo of Jeffrey Kiehl

"Humans will see a climate that our species has never experienced."

—Jeffrey Kiehl, NCAR

Some of the most profound episodes in the history of life on Earth are being profiled with new clarity by the NCAR Community Climate System Model. The CCSM boasts an interactive ocean, many types of vegetation, and other components that provide a more realistic take on the wrenching changes wrought by past climates.

In one major leap forward, the CCSM has simulated Earth’s recovery from the last ice age in unprecedented detail. Assisted by NSF support, supercomputers at the U.S. Department of Energy Oak Ridge National Laboratory racked up five million hours of processing time in 2009–10 for the TraCE-21000 project (Transient Simulation of Climate Evolution over the Last 21,000 Years). Directing the project are Bette Otto-Bliesner (NCAR) and Zhengyu Liu (University of Wisconsin–Madison).

With computer time at a premium, many paleoclimate studies span only a few slivers of geologic time. Others cover longer periods but cannot include as much detail. Thanks to the DOE resources, plus NSF support, TraCE-21000 managed to simulate its entire time span in full detail. The project enlisted a wide range of specialists from many disciplines and institutions to ensure the fidelity of its depictions. “It is a huge—and very successful—collaboration,” Liu says.

Photo of two men and one woman
Zhengyu Liu and John Kutzbach (University of Wisconsin-Madison) are collaborating with NCAR's Bette Otto-Bliesner on the TraCE-2100 project, one of the most ambitious simulations in the history of climate modeling.

NCAR’s Jeffrey Kiehl is exploring even more dramatic convulsions from Earth’s distant past that had dire consequences for many forms of life. “It’s the deep-time period that really interests me,” says Kiehl. He and colleague Christine Shields have analyzed the Great Dying, the extinction of more than half of all species on Earth about 251 million years ago. With CCSM, they explored whether the die-off resulted from a surge in volcanic activity that emitted massive amounts of carbon dioxide. They found that high-latitude oceans may have warmed so much and so deeply—as far down as 4,000 meters (12,000 feet)—that the normal downward flow of oxygen-rich surface waters would have been impeded. In this emerging paleoclimate, land creatures would have struggled with extreme heat while marine life was starved for oxygen.

Visualization in blue, red, white and black of Earth
The TraCE-21000 project is examining the Bølling-Allerød, Earth’s most recent period of natural global warming on a major scale, which took place about 14,500 years ago. Several factors combined then to raise temperatures dramatically over Greenland and parts of Eurasia (red and orange). (Courtesy Jamison Daniel, National Center for Computational Sciences, Oak Ridge National Laboratory.)

Working closely with geologists and glaciologists, Kiehl is now analyzing other time swaths, including the late Ordovician (about 445 million years ago) and the mid-Permian (about 260 million years ago). He’s deeply concerned about the pace of human-induced greenhouse gas release, which is far more rapid than any natural release in the geologic record. As he puts it: “We are performing a unique experiment with significant consequences for humanity.”