HILO, Hawai'i — In recent years the supernovae created by exploding white dwarf stars emerged as a kind of critical cosmic yardstick. They act as reliable beacons scientists can use to calculate distances in a mysterious, expanding universe.

Now astronomers using the telescopes on Mauna Kea on the Big Island have discovered those "Type 1a" supernovae have mysteries of their own that may make them less reliable as interstellar signposts.

An international team of researchers found a supernova, known as SNLS-03D3bb in a galaxy 4 billion light years away, that isn't supposed to be possible. It is far more massive than scientists had thought such a supernova could be, and it burns more brightly than any seen before.

"We have to figure out what this means, because there may well be another mechanism for producing supernovae that we haven't yet talked about," said Richard Ellis of the California Institute of Technology, division of physics, mathematics, and astronomy.

The discovery was based on images collected by the Canada France Hawai'i Telescope on Mauna Kea, and observations by Ellis using the W.M. Keck Telescope. The lead author of the study was University of Toronto postdoctoral researcher Andy Howell, and the findings were published in the journal Nature.

Type 1a supernovae burn with about the same brightness as one another, which has made them a "fundamental tool" for calculating distances and collecting other data, Ellis said.

The new findings are important because data gathered from observations of Type 1a supernovae were key to one of the greatest discoveries in astronomy, the astonishing 1998 finding that the expansion of the universe is speeding up instead of slowing down.

If the pieces of the universe are flying apart at ever increasing speeds, that meant something immense was missing from scientists' understanding of the forces at work. Theorists have been puzzling ever since over what that so-called "dark energy" could be, or what mysterious force could be causing that.

Ellis and the other researchers said their questions about the strange Type 1a supernovae they studied don't undermine the 1998 findings that the universe is expanding at an accelerating rate. Astronomers have studied 800 to 900 Type 1a supernovae, and this is the first to break the expected pattern, Ellis said.

But until the new findings can be explained, Ellis said they raise the possibility scientists could err in their calculations and conclusions because they misunderstand how the supernovae work.

"The thrust of this paper is really a warning that supernovae are very valuable objects for measuring the properties of the universe, but there are still things that we don't know," he said.

The supposed limit to the size of a white dwarf that explodes into a supernova was calculated in 1930 by Subramanyan Chandrasekhar, who won the Nobel Prize for using fundamental physical laws to demonstrate that such a white dwarf could not be larger than 1.4 times the mass of our sun.

The problem is the white dwarf that produced SNLS-03D3bb was apparently twice the mass of our sun, according to the new Nature article.

The thermonuclear explosions of supernovae produce the element nickel, and the astronomers studying the odd supernova observed that this particular explosion was producing far more of the element than had been seen in such an explosion before, Ellis said.

Astronomers also observed that the explosion was slower than any previously seen supernova, another indication that its mass was much more than had been previously thought possible.

One theory that might account for how the dwarf star became so large before it exploded into a supernova is that it was actually two white dwarfs that merged briefly before exploding.

Another theory is the original star may have been rotating so fast that centrifugal force kept gravity from crushing it as it does in other supernovae.

Reach Kevin Dayton at kdayton@honoluluadvertiser.com.