Extreme weather — including heat waves, floods and cold snaps — is likely to become more common during the next century in the United States because of human-generated greenhouse-gas emissions, according to a new study by Purdue University researchers.

The analysis, which is being published online this week in the Proceedings of the National Academies of Science, examines how heat-trapping gases linked to climate change may intensify precipitation, drought and other weather. Instances of extreme heat and cold would likely increase throughout the country, the scientists concluded, and many areas would likely experience heavier downpours even if rain became less frequent.

"I would be thrilled to be wrong," said Noah Diffenbaugh, a climate scientist at the Purdue Climate Change Research Center and the Department of Earth and Atmospheric Sciences. "It's definitely going to be more extreme hot temperatures."

The four-person research team, which included two scientists from the Earth Systems Physics Group at the Abdus Salam International Centre for Theoretical Physics in Trieste, Italy, compared U.S. weather patterns from 1961 to 1985 with models of future weather patterns from 2071 to 2095.

Under this scenario, which assumes the amount of carbon dioxide in the air would roughly double in the next 100 years, the coldest days of the year in the Northeast would be up to 18 degrees Fahrenheit warmer, and the temperatures currently experienced on the 18 hottest days of year in the Washington, D.C., area would prevail for two months.

The Southwest would become drier and hotter, the paper predicts, while the Gulf Coast would become warmer and experience less frequent, but more intense, rains.

Last month, three scientists from the National Center for Atmospheric Research (NCAR) in Boulder, Colo., published a paper suggesting precipitation throughout the world would likely become more intense as global warming continued. One of the authors of that study, Gerald Meehl, said the Purdue paper was consistent with his findings.

"This is what you'd expect to see," Meehl said. "As the climate continues to warm, you'd see more precipitation intensity [because] the warmer air can hold more moisture."

Kevin Trenberth, who heads NCAR's climate-analysis section, said the two findings highlight "the need to examine not just the precipitation amount, but also the intensity, frequency, duration and type" of rain or snow that may stem from climate change.

But Patrick Michaels, a senior fellow in environmental studies at the Cato Institute, questioned the Purdue paper because the scientists assumed carbon-dioxide concentrations would double.

"That's not going to happen," said Michaels, who has accepted funding from coal, gas and mining interests opposed to mandatory curbs on greenhouse-gas emissions. "These are very high assumptions that are not borne out by reality."

Diffenbaugh said his team did choose "a high-end scenario" in predicting greenhouse-gas emissions. He said it took five months to run the climate model on a supercomputer and said he and other scientists did not have the time to calculate the impact of less carbon-dioxide pollution.

Catastrophic weather has taken an increasingly heavy financial toll on U.S. homeowners and businesses.

In the past 30 years, the country has experienced a 15-fold increase in insured losses from extreme weather, according to a report last month by Ceres, a coalition of investors that lobbies businesses to be environmentally responsible. Increased development in flood-prone areas has contributed to those losses.

"We're in an era of escalating climate-change impacts on governments and the public that can cause substantial financial risks," said Ceres President Mindy Lubber.

"The government needs to step in and act, and call for policy changes that will reduce the risks."