The National Academy of Sciences, reaching that conclusion in a broad review of scientific work requested by the US Congress, reported on Thursday that the "recent warmth is unprecedented for at least the last 400 years and potentially the last several millennia".
A panel of top climate scientists told US legislators that the Earth is heating up and that "human activities are responsible for much of the recent warming".
Their 155-page report said average global surface temperatures in the northern hemisphere rose about one degree during the 20th century.
This is shown in boreholes, retreating glaciers and other evidence found in nature, said Gerald North, a geosciences professor at Texas A&M University who chaired the academy's panel.
The report was requested in November by the chairman of the House Science Committee, Republican Sherwood Boehlert, to address naysayers who question whether global warming is a major threat.
Boehlert said on Thursday the report shows the value of having scientists advise Congress.
"There is nothing in this report that should raise any doubts about the broad scientific consensus on global climate change," he said.
Other new research on Thursday showed that global warming produced about half of the extra hurricane-fuelled warmth in the North Atlantic in 2005, and natural cycles were a minor factor, according to Kevin Trenberth and Dennis Shea of the Commerce Department's National Centre for Atmospheric Research.
Their study is being published by the American Geophysical Union.
The Bush administration has maintained that the threat is not severe enough to warrant new pollution controls that the White House says would have cost five million Americans their jobs.
Climate scientists Michael Mann, Raymond Bradley and Malcolm Hughes had concluded the Northern Hemisphere was the warmest it has been in 2,000 years.
Their research was known as the "hockey-stick" graphic because it compared the sharp curve of the hockey blade to the recent upstick in temperatures and the stick's long shaft to centuries of previous climate stability.
The National Academy scientists concluded that the Mann-Bradley-Hughes research from the late 1990s was "likely" to be true, said John 'Mike' Wallace, an atmospheric sciences professor at the University of Washington and a panel member.
The conclusions from the 1990s research "are very close to being right" and are supported by even more recent data, Wallace said.
The panel looked at how other scientists reconstructed the Earth's temperatures going back thousands of years, before there was data from modern scientific instruments.
For all but the most recent 150 years, the academy scientists relied on "proxy" evidence from tree rings, corals, glaciers and ice cores, cave deposits, ocean and lake sediments, boreholes and other sources. They also examined indirect records such as paintings of glaciers in the Alps.
Combining that information gave the panel "a high level of confidence that the last few decades of the 20th century were warmer than any comparable period in the last 400 years," the academy said.
Overall, the panel agreed that the warming in the last few decades of the 20th century was unprecedented over the last 1,000 years, though relatively warm conditions persisted around the year 1000, followed by a Little Ice Age from about 1500 to 1850.
The scientists said they had less confidence in the evidence of temperatures before 1600. But they considered it reliable enough to conclude there were sharp spikes in carbon dioxide and methane, the two major "greenhouse" gases blamed for trapping heat in the atmosphere, beginning in the 20th century, after remaining fairly level for 12,000 years.
Between 1 AD and 1850, volcanic eruptions and solar fluctuations were the main causes of changes in greenhouse gas levels. But those temperature changes "were much less pronounced than the warming due to greenhouse gas" levels by pollution since the mid-19th century, it said.
The National Academy of Sciences is a private organisation chartered by Congress to advise the government of scientific matters.