The amount of floating sea ice in the Arctic Ocean acts as an important proxy for the overall global climate. With a greater area of sea ice coverage—what’s called its “extent”—the Arctic region stays cooler because more sunlight gets reflected back into space. When there is less sea ice, the ocean absorbs more solar energy, warming itself and the surrounding area.
Scientists have monitored Arctic sea ice extent reliably by satellite since the late 1970s and have used those data to help predict when the region will be completely free of summer ice. New statistical modeling from economists Francis X. Diebold of Penn and Glenn D. Rudebusch of the Federal Reserve Bank of San Francisco reveals this could happen sooner than previously projected, with a 60% chance of an effectively ice-free summer Arctic in the 2030s.
“Using statistical modeling, we’re predicting a much faster reduction of Arctic sea ice than most climate models have, by about 2040 as opposed to 2060 or 2070,” says Diebold, a professor of economics, finance, and statistics. “That’s strikingly sooner.”
Arctic surface temperatures are already increasing at least twice as quickly as global temperatures, and during the past 40 years about 40% of northern summer sea ice has disappeared. Such changes are an important barometer for the ways and pace at which the planet’s climate is shifting. “Moving forward, for the rest of the century and onward, it’s hard to think of anything more globally important,” Diebold says.
For the past 35 years, starting in graduate school, he and Rudebusch have worked together on a variety of subjects. About two years ago, they turned their attention to climate, spurred by the gravity of the unfolding situation and the realization that the tools of statistics and econometrics could be put to good use on these important issues.
“Climate change will change everything,” Rudebusch says. “If you’re an economist or businessperson and think climate change is something just for scientists, no. It’s going to have huge effects on the economy and financial risk and reward. For example, we will have to make substantial investments to decarbonize the economy.”
Current climate models analyze what is happening on the land, in the sea, and in the air at a very granular level in space and time. Those interconnections paint a crucial picture about climate variation over time, but climate models have tended to underestimate the trends in sea ice loss.
That’s where Diebold and Rudebusch felt they could contribute. Using the National Snow and Ice Data Center’s sea-ice-extent data from November 1978 through October 2019, the economists statistically analyzed the long-term future evolution of Arctic sea ice, paying close attention to time variation in seasonal patterns and trends. Their methodology also introduced something they called “shadow ice” to account for the fact that the total amount of sea ice can’t fall below zero.
They also calculated uncertainty ranges for their projections. This probability approach is essential because their forecasting used past data to predict a range of future scenarios. “All of our analysis is based on business as usual,” Rudebusch says. “But if the future changes drastically, then, of course, the probabilities will change. That being said, given the economic development, population, and emissions trends of the past, this is the future we envision for sea ice.”
The researchers say they hope these statistical results can help calibrate future climate models. “The melting Arctic is really a microcosm, intensified, of all that’s going on,” Diebold says. “It’s almost as if you can view the future by looking at the Arctic.”
Francis X. Diebold is the Paul F. and Warren S. Miller Professor of Social Sciences, a professor in the Department of Economics in the School of Arts & Sciences, and a professor in the Departments of Finance and Statistics in the Wharton School.
Glenn D. Rudebusch is executive vice president and senior policy advisor of monetary economics, macroeconomics, and finance at the Federal Reserve Bank of San Francisco.
