 Dr. Darko Koracin with DRI's new Origin2000 supercomputer. This powerful machine will benefit modelers working in the Division of Atmospheric Sciences |
DRI researchers Drs. Darko Koracin, Melanie Wetzel, and Steve Chai, along with DRI computer systems administrator Dr. Jeff Tomer, get the credit for bringing the $250,000 Origin2000 to DRI. With it they plan to make leaps and bounds in their research to predict and understand such atmospheric phenomena as coastal wind jet, cloud and fog formation and evolution, and the dispersion of pollutants in mountainous regions.
"Not many institutions have something like this," says Koracin. "It's a very powerful tool." The purchase was made possible by a $205,000 grant from the Defense University Research Initiative Program of the U.S. Department of Defense (DOD) combined with $50,000 from DRIıs Applied Research Initiative Fund, a program created by the Nevada legislature to support the development of research in the state.
The idea to get a supercomputer had been simmering for some time, explains Koracin. "As atmospheric modeling became more important here at DRI, there was a lot of discussion about how to equip the Institute better." Modeling uses numerical computer simulations to recreate and predict atmospheric processes. "It's crucial," says Koracin, "for applied research in areas like air quality and weather prediction, as well as for more fundamental research aimed at understanding the various underlying factors that control weather and climate."
The time was right to improve DRIıs ability to compete in this area, and the partnership with DOD was a natural. Koracin, Chai, and Wetzel are all conducting research for DODıs Offce of Naval Research that will be enhanced by this computer's capabilities. The new capabilities also will further existing collaborations that these researchers have with the Naval Postgraduate School and Naval Research Laboratory in Monterey, and the Scripps Institute of Oceanography in San Diego.
And those capabilities are quite impressive. Technophobes may shrug, but those who know computers beyond the mouse and keyboard will be impressed to hear that the Origin2000 has no less than 16 300-megahertz microprocessors (a standard PC has just one), two gigabytes of RAM or short-term memory (equivalent to about 50 PCs), and approximately 100 gigabytes of hard disk capacity. It's those 16 processors that really set the Origin2000 apart, says Koracin. "The multiple processors are the biggest advantage. It could take months, or be completely impossible, to run some of these models on a machine with just one."
The need for this kind of computing power all comes down to the details. In complex terrain like the mountainous West, or in coastal areas, there are more factorstopography, ocean temperatures, sea breezesthat can affect airŝow and weather patterns. For atmospheric models of these areas to be accurate and useful for predictions, they must have a very high resolution, or incorporate a very detailed account of those various features. Large-scale models like those used in media forecasts are crude in comparison, concerned primarily with the general movement of large fronts while ignoring the localized features that may affect weather in a speciŜc area. "That's why the Navy, for instance, canıt use global models to predict weather in a coastal zone," explains Koracin.
"The Origin2000 gives us a pretty good edge," says Koracin. "We can offer more detailed models, extend ones we are using currently, and propose some new avenues we havenıt been able to explore before."
For instance, Koracin will enhance his work for DOD's Offce of Naval Research, where he is using several models to understand how the cloudy marine atmosphere off California's Monterey coast responds to factors like coastal mountains and changes in ocean temperatures. The faster computer means that higher resolution models can be run over longer periods of time, increasing the accuracy of predictions for this atmospherically complex and changeable region.
Other DOD projects slated to get a boost from the Origin2000 include work by Chai and Wetzel using satellite images to characterize clouds and relate them to a marine forecasting model recently developed by the OfŜce of Naval Research. Another study, led by Koracin, involves the ŝow of air currents over rough or uneven land surfaces, and the impact of these currents on the transport and dispersion of atmospheric pollutants.
This type of research could have important applications in the mountainous West which has more than its fair share of "rough and uneven" land surfaces. In fact, Koracin says that the Origin2000 has great potential for improving air quality through application of dispersion models. High resolution models can help predict where and how pollution from an urban area, say Las Vegas, will spread. "We already have an air quality study in Las Vegas. I would like to see how we can extend and improve that project with the help of this computer." Koracin also plans to approach the Washoe District Health Department about creating an operational model for the Reno area, where pollution problems are often closely linked to weather phenomena such as winter inversions.
Koracin, Wetzel, and Chai are educators as well as researchers (each teaches graduate-level courses at the University of Nevada, Reno and employs students as research assistants). The Origin2000 will, therefore, serve as a teaching tool, helping to train the next generation of scientists. In the meantime, the current generation of scientists will have the opportunity to make some lasting contributions to the understanding of atmospheric processes and the mitigation of air quality problems.
Jackie Allen and Roger Kreidberg
|
|