Experts who have been poring over data and commercial satellite imagery now say they believe the yield of North Korea's Sept. 3 nuclear test was significantly greater than initially estimated.

An analysis by the respected 38 North website, which is published under the auspices of the U.S.-Korea Institute at the Johns Hopkins School of Advanced International Studies, concludes the blast was as powerful as 250 kilotons, or one-quarter megaton.

That puts the test clearly in the range of what would be expected from a powerful hydrogen bomb, and is significantly higher than what U.S. officials have confirmed.

It's also more than 12 times as powerful as North Korea's previous test conducted in September 2016, which was calculated to have a yield of 10 to 20 kilotons, and 17 times as big as the 15-kiloton bomb the U.S. dropped over Hiroshima in 1945.

Defense Secretary Jim Mattis, speaking to reporters traveling with him on a tour of the U.S. nuclear forces at Minot Air Force Base, N.D., Wednesday, refused to say if the North Korean test was of a two-stage hydrogen bomb, or reveal its yield, even though he said the U.S. has "a pretty good idea of what happened."

Mattis said he does not dispute published accounts of the bomb's yield, which range from a low of 50 kilotons reported by South Korea to Japan's 160 kilotons.

"100 kilotons or more," Mattis said. "So it's a large one. I don't want to talk any further than that right now, OK? It's a large one."

The 38 North site initially calculated the blast at 120 kilotons, but a new analysis by a trio of nuclear scientists now puts the yield at more than double that, based on the latest seismic readings from United States Geological Survey, which put the magnitude of the resulting earthquake at 6.1.

"This revision is significant because, rather than providing an equivalent yield of about 120 kilotons derived from the lower magnitude estimates, the application of standard formula with appropriate constants shows that the yield can now be estimated to have been roughly 250 kilotons," the scientists write. "This large explosive yield is also quite close to what 38 North had previously determined to be the maximum estimated containable yield for the Punggye-ri test site."

The scientists downplayed the risk that the North Korean mountain where the test was conducted might collapse releasing radioactive gas.

"Such reports of the mountain's demise are highly exaggerated," the scientists said, "although there is little doubt that there would have been significant ‘cracking' (possibly extending to the surface) as well as significant 'irreversible strain' resulting from this event."

The analysis also notes that North Korea still has two unused tunnel complexes it could use for future nuclear tests.