By Madeline Stone, National Geographic
A database created in part from 19th-century maritime records sharpens our view of climate change over the past 150 years.
In September of 1879, the Arctic-exploring USS Jeanette was sailing north of the Bering Strait when it was surrounded by ice floes and frozen in place. Imprisoned at sea, the 33-person crew struggled to survive for nearly two years before their ship sank, forcing them to embark on a perilous journey back to civilization. While they were stranded, the crew took down regular observations of the weather—winds, clouds, air pressure, temperature—creating a detailed meteorological record where no others existed.
One hundred and forty years later, that record is now helping scientists reconstruct Earth’s weather and climate history in unprecedented detail.
The USS Jeanette’s logs, which eventually made their way back to the United States along with 13 haggard crewmen led by chief engineer George Melville, were among the very first to be rescued as part of the Old Weather: Arctic project, a citizen science-fueled effort to digitize and transcribe the weather observations made by U.S. military vessels that sailed the Arctic in the 19th and 20th centuries. Those records, along with similar data housed in many other archives, are being fed into the 20th Century Reanalysis, a sophisticated weather reconstruction database developed by the National Oceanic Atmospheric Administration that allows scientists to characterize floods, droughts, storms, and other extreme events from history—and use the violent weather of the past to understand the present.
Earlier this month, that reconstruction received a major update when scientists infused it with millions of new observations from old ship’s logs and weather stations around the world. Now, NOAA’s souped-up “weather time machine” can produce snapshots of Earth’s atmosphere eight times a day going all the way back to 1836.
“Every three hours, we’re providing an estimate of what the weather was, anywhere in the world,” says Laura Slivinski, a research scientist at the University of Colorado Boulder’s Cooperative Institute for Research in Environmental Sciences (CIRES) and NOAA’s Earth System Research Laboratory. “It’s pretty unique.”
The ‘fog of ignorance’
Today, scientists have myriad satellites and weather stations at their disposal to study the weather. But satellite record keeping only began about 40 years ago, and prior to the mid-20th century there were far fewer weather stations. Scientists can use models to “hindcast” the weather further back in time, but without data to feed into those models, their reconstructions are murky.
“We call it the fog of ignorance,” says Gilbert Compo, a senior research scientist at NOAA’s CIRES.
To cut through that fog, researchers at NOAA have spent more than a decade gathering data on surface pressure, temperature, and sea ice conditions from archives around the world that are being digitized and transcribed with the help of volunteers. These data rescue efforts include several iterations of the Old Weather project, a project that digitized hand-written weather reports from 19th-century England, one focused on log books kept by Australian sea captains, and many more.
Once written records have been placed in a format NOAA can use, they can be added to the 20th Century Reanalysis, which uses a model similar to the one the National Weather Service relies on to make forecasts to produce snapshots of the atmosphere going back in time. The latest version of the reanalysis includes 25 percent more observations for years prior to 1930, resulting in far more reliable hindcasts, particularly for the 19th century.
Every additional ship’s log helps. For instance, in October 1880, a famously powerful cyclone made landfall in the town of Sitka in the Alaska panhandle. Older versions of the 20th Century Reanalysis couldn’t recreate this storm at all. But the latest update includes observations from the USS Jamestown, a vessel that was moored offshore at the time. With its pressure readings, the weather time machine is now able to produce a storm in the right location at the right time.
“It feels like a drop in the bucket, but those observations add up,” Slivinski says.