New ideas about Earth's formation

Scientists have long wondered why the contents of Earth's atmosphere doesn't match the contents of the rest of the solar system. Late last year, Lawrence Livermore scientists made a discovery that may lead to a new model of how Earth was formed and may, in turn, explain this puzzle.
Livermore's Marc Caffee and his team described in the September 24, 1999, issue of Science their discovery of xenon gas seeping out of gas wells in Colorado, New Mexico, and Australia. The gas is of the same form as xenon found in the sun and throughout the solar system, but different from that found in Earth's atmosphere.
The discovery hints at a more complex picture of how Earth's atmosphere evolved. Scientists have long posited that Earth and the rest of the solar system formed about 4.5 billion years ago from a disk of dust swirling around the newly born sun. Earth formed itself into separate layers-core, mantle, and crust. Gases bubbled out of the mantle to create an atmosphere around the planet.
The discovery of primordial xenon gases suggests that "the old paradigm of one big happy uniform mantle, and the atmosphere coming out of all or part of it, is too simple," said Caffee. If the gases, which are thought to come from the mantle, are similar to what is found in meteorites but not to what's found in Earth's atmosphere, did something happen to Earth to alter its atmosphere? Or is it that the mantle is much more complex than previously thought?
One notion is that the early Earth may have been put together from a plethora of materials that left different chemical signatures in the mantle. Earth's atmosphere may have come from only a part of the mantle, a part that did not include the primordial gases.
At any rate, those gases provide a view of the planet's earliest days. "It's kind of cool that the stuff is down there," said Caffee.
Contact: Marc Caffee (925) 423-8395 (caffee1@llnl.gov).

A primer on parallel computing

The newly published Industrial Strength Parallel Computing (Morgan Kaufmann Publishers, November 1999), edited by Livermore's Alice Koniges, sounds daunting. But the book really aims to make software developments in parallel computing more accessible to the growing numbers of parallel computing practitioners.
Back in the 1960s, parallel computing existed but wasn't much applied. In changing computer processing from step-by-step sequential calculations to independent but simultaneous calculations of many parts of a large problem, the new technique demanded increased processing power and new software to handle a new kind of problem management.
In the beginning, parallel computing was largely relegated to research universities and national laboratories. The introduction of PCs, which gave more people access to many inexpensive processors, pushed more researchers to develop software for parallel computing.
Among the efforts to enhance software developed by national laboratories and universities for industrial applications was the Parallel Applications Technology Project, a collaborative effort funded by the Department of Energy, similar agencies worldwide, Cray Research, and industrial partners. The book edited by Koniges grew out of the project when its leaders thought their results "needed to be collected in more than a series of articles in scientific journals."
Koniges says the book represents the expertise of some 72 contributors from national laboratories, universities, and private industry. Eleven of its 25 chapters have Livermore authors. Livermore contributors include Steve Ashby, Chuck Baldwin, William Bosl, David Eder, Robert Falgout, Morris Jette, Douglas Rotman, Steven G. Smith, Vijay Sonnad, John Tannahill, Andrew Thompson, and Lin Yang.
Koniges has spent most of her 15 years at Livermore in developing leading-edge scientific computing technology. She is an internationally recognized authority on parallel applications development.
Contact: Alice E. Koniges (925) 423-7890 (koniges1@llnl.gov).

Back to March 2000