New dataset provides the most complete look yet at the climate of the last 2000 years

The natural world holds a wealth of information about how hot or cool the planet was in the past – from records stored in tree rings and tropical corals to clues hidden in glacier ice and tiny fossils at the bottom of the ocean. Now, a new database taps into that information, creating the most complete timeline of past shifts in the globe’s temperatures over the last 2000 years. The effort, which was published today in Scientific Data, includes 692 records of historic climate change from 648 sites around the world.

It is the brainchild of the 2k Consortium, a collaboration sponsored by the Future Earth global research project Past Global Changes (PAGES). The database, which is open access and available online, is more than just big: The research team used rigorous screening methods to ensure that the data they captured was an accurate reflection of Earth’s past climate. Julien Emile-Geay, who helped to lead the development of the new database, says that it will be an important tool for showing how strange recent decades have been in the history of the planet.

“Because of manmade climate change, the 20^th Century is something that every climate scientist knows at some level to be unusual,” says Emile-Geay, an associate professor of Earth sciences at the University of Southern California in Los Angeles, United States. “This is going to be an incredibly clean resource for addressing such questions more rigorously.”

The maelstrom

Records of past climate conditions on Earth, what scientists call “paleoclimate” data, aren’t direct measurements of temperature – no thermometers existed in the Middle Ages. Scientists, instead, make inferences from the information they find in the natural world, says Nicholas McKay, an assistant professor of climate science at Northern Arizona University in Flagstaff, United States.

A climate scientist, for example, might travel by boat to the ocean, pull up sediment from the bottom, and take it back to the lab to make measurements. “But that thing I’m measuring is not temperature,” says McKay, who led the effort to create the new database with Emile-Geay. “It's the abundance of some kind of bug. It's the absorption of some wavelength of light. Then there’s some process that goes from that measurement to me saying I think the variation in that quantity reflects variation in past temperature.”

In once case of data collected from marine sediment, the shells of microscopic organisms called foraminifera take on different chemical compositions depending on how warm the surrounding water is. Scientists can analyse fossils of these organisms to estimate the temperature hundreds of years ago. Trees, in contrast, often grow more in warmer years than in cooler years, creating rings of different widths depending on the weather they encounter.

That can, however, make paleoclimate data difficult to work with. Across the globe, hundreds of different teams are collecting natural records of climate, and they all take different approaches to storing their data – meaning that, to get a full picture of Earth’s historic temperatures, you have to navigate a “big maelstrom” of datasets, Emile-Geay says.

To get around that chaos, he and his colleagues opted to put all that data into one place and in the same format. They published their first mega-dataset in 2013, but the latest effort, which was produced by 98 scientists from 22 countries, is more complete. It includes, for example, 415 records from tree rings, 96 from coral, 58 from marine sediments and more. When put together, these records can support estimates of global temperatures for every year dating back two millennia.

The team also added in something that scientists call “metadata” – which aren’t measurements but explanations of how scientists obtained their measurements and what those measurements mean. Such information is critical, McKay says, because it can help scientists to screen out paleoclimate records that don’t have a strong link to temperature. Depending on where they grow, some trees, for example, record less about temperature and more about moisture in the soil. That means that their rings don't provide accurate estimates of temperature. To weed out those sorts of records, McKay and his colleagues collaborated with experts in paleoclimate archives from various regions of the world and in different types of data.

The first database of this kind sponsored by PAGES, McKay says, could fit in a single Microsoft Excel spreadsheet. “But there’s so much other information that you can’t squeeze into an Excel spreadsheet,” he says. “In this version, we developed a much richer system, so we could include all those rich metadata, including how the developers of each record interpret their data and how it relates to temperature.”

Hockey stick

The new database is useful for more than making the lives of scientists easier, Emile-Geay says. It provides an invaluable window to Earth’s past, which, he says, scientists must understand if they want to know how the planet will change in the future.

“One of the exciting parts about this is that we honestly don’t know all the ways that people are going to use this database,” Emile-Geay says. “It was by design that we created something that was as complete as possible to enable all new lines of scientific inquiry.”

In one example, he and his colleagues took all of the datasets that they had gathered and stacked them on top of each other. This simple approach – which Emile-Geay says avoided the sophisticated statistics that climate skeptics often rail against – showed a familiar pattern: Temperatures remained mostly flat for 2000 years, then shot up in the last few decades.

“Lo and behold, the pattern we got what looks like a hockey stick, which is, of course, not entirely surprising if you’ve been following this field for the last 20 years,” Emile-Geay says. But given the heated pushback against climate science in recent years, “it’s nice to see that this broad-scale pattern has stood the test of time and appears over and over again no matter how you treat the data.”

He and McKay both say that one of the highlights of the work was the international collaboration. Getting close to 100 scientists from different cultures, both national and scientific, to work together wasn’t easy, but it was worth it.

“It was just such a beast of an effort in that sense,” McKay says. “On the other hand, it was incredibly valuable to have this perspective. Because it’s so hard to get scientists to agree on anything, when you can get a large body to agree, it ends up being a sound, robust product, one that we hope will be an enduring legacy for the field as a whole."

To find out more about the PAGES 2k Network and its projects, go to: pastglobalchanges.org/ini/wg/2k-network/intro