The meteorite that wiped out the dinosaurs might have been less than half the size of what previous models predicted.

That's the finding of a new technique being developed to estimate the size of ancient impactors that left little or no remaining physical evidence of themselves after they collided with Earth.

Scientists working on the technique used chemical signatures in seawater and ocean sediments to study the dino-killing impact that occurred at the end of the Cretaceous period, about 65 million years ago.

They also looked at two impact events at the end of the Eocene epoch, roughly 33.9 million years ago.

In what could be a major scientific puzzle, the team's new size estimate for the dino-killing meteorite is a mere 2.5 to 3.7 miles (4 to 6 kilometers) across.

The most recent computer models predicted a size of 9 to 12 miles (15 to 19 kilometers) across.

The team notes that their findings could also mean that the makeup of the impactor is different from what scientists commonly assume.

"We are hoping this will lead to further work," said study leader Gregory Ravizza of the University of Hawaii in Honolulu.

Impact Fingerprints

The fiery passage of asteroids and comets through Earth's atmosphere leaves chemical traces in the land, sea, and air.

The most common types of meteorites to hit Earth are chondrites, stony objects that originate in the asteroid belt.

Chondrites contain two different versions, or isotopes, of the naturally occurring element osmium: osmium 187 and osmium 188.

Seawater and sediments also contain the two osmium isotopes, but the ratio of osmium 187 to osmium 188 is usually much larger in the ocean than it is in chondrites.

When a small- to medium-size meteorite enters Earth's atmosphere, much of the object is vaporized and the osmium ratio in seawater around the world is temporarily decreased.

Over time, this osmium imprint is transferred to sediments at the ocean bottom, creating a more enduring record of the impact.

The new technique therefore looks for osmium spikes in ocean sediments and analyzes the isotope ratio. Scientists can then predict when an impact event occurred and the size of the projectile.

The research is detailed in tomorrow's issue of the journal Science.

Dramatic Upheaval

In addition to the smaller Cretaceous impact, the team estimates that two known meteorites from the late Eocene were smaller than previously believed.

Boris Ivanov, an impact modeler at the Russian Academy of Sciences, said that if the new size estimates prove correct, they would create a "dramatic controversy" within the impact physics community.

"Most numerical modeling specialists believe the current modeling gives us fidelity of a factor of a few times the mass of a projectile with assumed average impact velocity," Ivanov said.

Study co-author Francois Paquay, also at the University of Hawaii, said that more work needs to be done to confirm the latest estimates.

"We think the discrepancy is important and it will need to be addressed in future [scientific] meetings," Paquay said.

Jay Melosh, a planetary scientist at Arizona State University who was not involved in the study, called the new method a "potentially powerful" technique for filling gaps in the geologic impact record.

"It's a very valuable contribution to the tool kit of ways we have of estimating the presence of impacts in the geologic record," Melosh said.