Sketchy data. The Times Atlas (left) shows 15% more glacier melt than scientists believe. A map drawn up by SPRI scientists (right) suggests that the map was erroneously based on ice thickness
Credit: (left) The Times Atlas of the World; (right) Toby Benham
CAMBRIDGE, UNITED KINGDOM—So much for claims that climate scientists deliberately misrepresent their data: glaciologists are broadly and loudly
panning the latest version of The Times Comprehensive Atlas of the World, released last week, which shows Greenland having lost 15% of its ice cover in the past 12 years due to warming, turning an area the size of the
United Kingdom and Ireland "green." The atlas is published by HarperCollins on behalf of London's The Times newspaper.
The trouble, researchers say, is that although Greenland's ice sheet is retreating, the melt is nothing like the scale shown in the atlas and they are
mystified at where the error arose. In a letter sent to HarperCollins on Friday evening, researchers at the Cambridge-based Scott Polar Research
Institute (SPRI) quickly attempted to set the record straight. "A sizable portion of the area mapped as ice-free in the Atlas is clearly still
ice-covered," they wrote. "There is to our knowledge no support for this claim in the published scientific literature."
"It's a really bad mapping error," glaciologist Liz Morris of SPRI told ScienceInsider. If 15% of ice was lost, then sea levels would have risen
by 1 meter. "That obviously hasn't happened," she says. "Most people with a science background would have spotted something wrong." While satellite
images show that ice in Greenland is certainly retreating in a way that is "very interesting and dramatic," those retreat patterns are far too small to
show in a map the resolution of the one in The Times Atlas. The 15% retreat, SPRI glaciologists have worked out, is 150 times the amount of ice
loss that has actually occurred.
On the glaciology listserv Cryolist, experts have been trying to work out the source of the error.
Their best guess so far is that the cartographers were measuring ice thickness, rather than the actual height of ice. As ice sheets are thinner where
submerged mountains exist, that would explain why the shape of the ice sheet is "bizarre," Morris says.
Another possibility is that from the air, ice sheets appear dark and are difficult to distinguish from the ground, whereas snow gives off a glare, says
geographer Graham Cogley at the Trent University in Canada. The extent of the error, he says, "sticks out like a sore thumb." The global average of
glacier melt is about 0.2% per year, and that includes very small glaciers, which melt more quickly than the massive Greenland ice sheet. A 1.5% per
year decrease in ice cover, he says, is "so implausible" that any glaciologist would spot it.
Cogley compares this mistake to a case where scientists were reported as predicting in 1999 that the Himalayan glaciers would melt by 2035, rather than
2350. This error was propagated through the popular science press and made its way into an Intergovernmental Panel on Climate Change (IPCC) report in
2007. Cogley tracked down its source in 2009 and the IPCC quickly retracted that part of its report—not, however, before climate change skeptics had
publicized the error.
Morris says that scientists are worried about skeptics picking up this error too as evidence of scientists exaggerating the effects of global warming.
"This was an error on the part of a mapper, not a glaciologist, and was compounded by the publicity department. As soon as scientists saw this, there
was absolute outrage," she says.
The letter to The Times read: "We do not disagree with the statement that climate is changing and that the Greenland Ice Sheet is
affected by this. It is, however, crucial to report climate change and its impact accurately and to back bold statements with concrete and correct
"It's fair to say that The Times book is the most authoritative," says Cogley "That's why it's such a pity they've blundered on this
point. …. I hope they're busily trying to figure out where they went wrong."
A spokesperson for HarperCollins, told the BBC that the data came from the U.S.
National Snow and Ice Data Center (NSIDC), adding that the 15% retreat since the 10th edition of the atlas was released in 1999 is a result of global warming
and "much more accurate data."
Cogley worries about that statement, and says the atlas data is definitely not accurate. "They might be about to make the mistake IPCC didn't … and
tough it out" rather than publishing a retraction, he says. "That pig won't fly."
"We're really nervous about this one; it's such a stupid mistake," says Morris. "But people can say scientists say [this], and it's just not true."
UPDATE, 2:25 p.m.:
Ted Scambos, lead scientist for NSIDC's science team in Boulder, Colorado, says that researchers have tracked down the probable source of the error: a
map NSIDC published in 2001 that showed the extent of Greenland's central, thickest ice sheet. It does not, however, show any of the peripheral
glaciers. This map, he says, stands on its own, but the atlas cartographers most likely took it out of context.
"It's unfortunate NSIDC wasn't contacted," he says. "Folks here respond quickly; not only could NSIDC have helped, but any number of groups could
instantly have known" that something was wrong. Now glaciologists are left trying to figure out how not to understate the importance of the extent glacial
ice melt, while at the same time correcting the error.
Sheena Barclay, managing director of Collins Bartholomew, which publishes The Times Atlas, says that she is waiting to hear back from The Times Atlas' editor, Jethro Lennox, but told ScienceInsider that it would be "unlike us not to speak to [the scientists at NSIDC] and
corroborate" the findings. But the publisher is standing by the 15% number, which she says comes from comparing "like to like" data between 1999 and
2011, and accounting for more accurate methods during that time. The atlas' introduction, she says, explains the issue of ice cap melt and highlights
that more research needs to be done.