NASA Climate Modeling Spurs Next Computing Revolution

Advanced computing systems are the only tools we have to model our planet as a whole interactive system to mimic nature's behavior. NASA's Earth Science Enterprise is dedicated to developing computer models that will unlock the secrets to how natural and human-induced changes impact our global environment.

Unlike atoms and cells, which can be studied in the laboratory, the only effective way to study the inner workings and future course of our environment is through advanced computer modeling capabilities, which will show global climate change through computation of massive and complex mathematical relationships.

"Through math and science, and the advancement of our computer capabilities, we can unlock the mysteries of our planet's life cycle," said Dr. Ghassem Asrar, Associate Administrator for the Office of Earth Sciences, NASA Headquarters, Washington, DC.

"Currently, we model the climate system in a degree by degree, latitudinal and longitudinal grids, allowing us to examine global and continental atmospheric and oceanic conditions," Asrar said. "That resolution lets us see what is happening in the current and near term climate system, but not in the long run and why change is happening, and the consequences of such change at local levels."

"Today we're announcing the selection of nine proposals that span across all of NASA that we expect to mature into advanced computing systems that will be robust enough to handle massive amounts of data every second -- the kind of platforms that will be able to incorporate vast amounts of Earth Sciences data in 'living' models of our global climate, yet able to resolve regional and perhaps local phenomena such as severe storms and hurricanes," Asrar added.

"These computer models will incorporate factors such as chemistry of the atmosphere and the physics of clouds and the variability in the Sun's radiation that reaches Earth. Increasing our computer capabilities into an advanced interface that allows for multi-discipline scientific models to operate together in a coherent, interoperable computing environment is the only way we can achieve the nation's goal of understanding long-term global climate change. The selection of these proposals will take us down the path toward improving both weather and climate predictive capabilities by three to five times over our current computing power.

"We need to make a leap from today's segmented and evolutionary systems to a unified, revolutionary pathway into the future of advanced computing," Asrar concluded.

NASA's efforts initiated here constitute a direct response to recommendations made recently by the National Research Council (NRC) in its report Improving the Effectiveness of U.S. Climate Modeling. In this study, the NRC states that efforts in climate modeling need to be linked together and with related efforts in the broader research community through a common infrastructure. The selected efforts, which constitute the combined efforts of scientists within NASA, other government agencies and the academic community, should greatly facilitate this needed component of the overall effort.

The connection between NASA's efforts and those of the other federal agencies that constitute the U.S. Global Change Research Program also is crucial to the improvement of the nation's climate modeling capability. A recent USGCRP report, "High-End Climate Science: Development of Modeling and Related Computing Capabilities," also made the argument for the development of a software infrastructure to support climate research such as has been initiated here by NASA.

NASA has selected nine proposals in response to an Agency-wide solicitation titled "Increasing Interoperability and Performance of Grand Challenges Applications in the Earth, Space, Life and Microgravity Sciences," announced in 2000. The proposals selected will now have their costs negotiated, with science teams expected to finish a joint-framework policy by 2003.

NASA has committed $18 million to this development phase, with the expectation that the first advanced Earth System Models will begin production in 2004. Additional tasks selected in response to this announcement cover a broad variety of topics in Earth and space science. NASA will subsequently announce individual team selections and award values this fall.