Faulty Uzbekistan Data Undermines Dire Forecast in Climate Study

A widely publicized climate study, which predicted severe global economic losses due to climate change, is now under scrutiny after a critical data error involving Uzbekistan was discovered, The Caspian Post reports citing foreign media.

As reported by The Washington Post on August 7, the error significantly skewed the study’s projections, prompting renewed debate over the reliability of economic modeling in climate science.

Published in Nature in 2023, the original study warned that unchecked climate change could reduce global GDP by 19% by 2050 and by an alarming 62% by 2100, nearly three times higher than earlier forecasts. The study attracted substantial media attention and became the second-most-cited climate paper in 2024, according to CarbonBrief. Its projections were used by U.S. government agencies and the World Bank in financial planning.

However, a new commentary in Nature, led by Solomon Hsiang, director of Stanford University’s Global Policy Laboratory, revealed that the study’s extreme forecasts were largely driven by distorted GDP data from Uzbekistan. Once researchers excluded Uzbekistan from the model, the projected global GDP losses dropped sharply from 62% to 23% by 2100, and from 19% to 6% by 2050.

The flawed dataset suggested that Uzbekistan’s GDP plummeted by nearly 90% in 2000, then rebounded by over 90% in certain regions by 2010, figures inconsistent with historical records. According to the World Bank, Uzbekistan’s actual annual growth between 1980 and 2020 ranged between -0.2% and +7.7%.

“These extreme swings warped the study’s model, creating the illusion that global GDP was far more sensitive to climate than it really is,” Hsiang told The Washington Post.

The original study’s authors, based at Germany’s Potsdam Institute for Climate Impact Research, acknowledged the error but stood by their conclusions. After revising the Uzbekistan data and adjusting the model, they reduced their 2050 forecast from a 19% to a 17% GDP loss.

“We’re grateful for the scrutiny,” said co-author Leonie Wenz of the Technical University of Berlin. “But the main conclusions still hold.”

Still, critics argue that retroactive methodological adjustments raise concerns about scientific integrity. “Science doesn’t work by adjusting experiments to get the answer you want,” Hsiang cautioned.

The incident highlights both the power and the fragility of large-scale climate modeling and the importance of validating every data point, regardless of a country’s size.