The goal of the Center for Integrated Space Weather Modeling (CISM) is to build a physics-based model of space weather capable of forecasting what will occur in the space environment near earth as a result of what happens on the sun. Just as hurricanes, drought, or even a heavy snowfall will affect people’s economic well-being—or simply force changes in their daily routine—so unusual or extreme space weather, despite its seeming remoteness, can have a similar impact. In May, 1998, for example, about 20,000 miles from the earth an "unusual peak in the fluence of relativistic electrons" caused a satellite in geosynchronous orbit to fail. Eighty percent of the pager systems in the United States were disrupted.
Having the ability to predict such events today could save society hundreds of millions of dollars a year. As the modern world becomes ever more dependent on technology that can be vulnerable to changes in the near-earth environment, the need for accurate forecasting only increases. At present, scientists’ capability for predicting space weather events is at best comparable to earth weather forecasting of a half century ago.
Reliable forecasting of any kind requires an accurate model of how the system whose behavior is to be predicted actually works. To be able to construct such a model, scientists need to know what underlying physical principles are at work in the system and how all of the system’s separate elements originate, develop over time, and interact.
A very unsophisticated "model" for predicting how long it would take matter thrown out from the sun to reach the earth, for example, is a simple formula: divide the distance between the sun and the earth by how fast the matter is travelling. Basic arithmetic is sufficient to "run" such a simple model.
However, the design of a model that could actually produce a useful forecast about something like space weather quickly becomes complicated as more and more contributing physical processes and other factors get included. In the case of matter thrown out from the sun, for example, such factors as: where the earth is in relation to the sun at the time the matter is thrown out, the mass of the matter, what the matter is made of, what volume of space it occupies, the strength of its electrical charge, what its magnetic field is, and so on—as well as the conditions the matter would encounter between the sun and earth and the conditions immediately around the earth when—or if—the matter does arrive there.
As researchers develop their model they test it by "predicting" an event that has already happened and about which they therefore have extensive information. Not all (and perhaps not any) of the model’s predictions will match the particulars of an actual event as it was observed. These "errors" or inaccurate predictions, however, actually are very useful to the further development of the model. They point researchers to those aspects of a phenomenon about which they need to know more. And they can also suggest avenues of investigation along which new discoveries of basic physics might be made.
Space weather is especially complex. It is a system of systems, made up of many phenomena that are controlled by different physical processes. At the core of the CISM model will be simulations of the four distinct regions between the sun and the earth where space weather takes place: the region immediately around the sun, particularly the solar corona; the region between the sun and the earth through which passes the solar wind; earth’s near-space environment, the magnetosphere; and the earth’s ionosphere and upper atmosphere.
Each of these regions is itself a system consisting of many interacting phenomena. The solar corona region alone, for example, is at least as complex as earth’s entire atmosphere. Yet a space weather model of the solar corona also has to include in its programming the origination and propagation of coronal mass ejections, solar flares, coronal holes, and solar prominences, all of which contribute to the solar wind. Coronal mass ejections and solar flares are also the origin of bursts of highly energetic electromagnetic radiation such as X-rays and gamma rays that also can affect the near-earth environment.
At present, computer models do exist for the four regions where space weather originates, develops, and interacts with the earth, but the models do not yet include all the physical processes involved. Nor are the models integrated: able to work together, to interact with each other as the phenomena actually do in space. The models are based on the physical principles that underlie the processes common to a specific region—and consequently, in their present state of development, they are limited to simulating only that region.
CISM’s first effort will be to integrate, for the first time, existing simulations of the four regions of space weather activity. This integration will be based on models already developed by Center members: a model of the solar corona developed by Science Applications International Corporation (SAIC), a model of the solar wind developed at the University of Colorado, a Dartmouth model of the magnetosphere, and a National Center for Atmospheric Research (NCAR) model of the interaction between the magnetosphere and earth’s ionosphere. New computer code must be written that will allow the outer boundaries of these four regions—what at present are mathematical limits—to be able to become fluid and active, able to translate from the mathematical conditions prescribed for one region into those of another.
Research will then focus on two goals: first, to continually modify or even redesign new models as the researchers’ understanding of the fundamental physics at work increases; and second, to include more of the complexity of specific phenomena being modeled. For example, separate grids or sub-models that simulate the rapid evolution of magnetic fields in active regions on the sun—which can generate solar flares and coronal mass ejections—are being developed by Center-member the University of California at Berkeley. These sub-models eventually will be incorporated into the SAIC solar corona model.
Similarly, in other first steps toward complexity, Center-member Rice University’s "Rice Convection Model" of the hot plasma in the middle magnetosphere will be nested within the Dartmouth magnetosphere model. And current work being done at several institutions on the transfer of matter and energy between the magnetosphere and the ionosphere will be used in working out the integration of the Dartmouth magnetosphere model and the NCAR ionosphere and upper atmosphere model.
A model having such complexity, where its major sub-systems are themselves composed of multiple interacting sub-systems, requires a massive computer capability. CISM will make use of the resources of the National Centers for Supercomputer Applications. At its home institution it will also have the use of one of the most powerful academic computer facilities not part of the National Centers: the high performance facility administered by Boston University’s Center for Computational Science and the Scientific Computing and Visualization Group. Having recently added three of the latest generation IBM eServer supercomputers, BU’s current computer capability is nearly a trillion calculations per second.
A parallel objective of the Center for Integrated Space Weather Modeling is to produce a dynamic visualization of its model. The visualization will be especially useful in producing forecasts and for teaching and education purposes. It also will be useful for researchers. The ability to actually see phenomena that are still only partially understood opens opportunities for new insight. In the three-dimensional virtual reality environment the Center’s visualization will create, a researcher will be able to "walk around" an event and see the development—in real or accelerated time—of otherwise invisible processes. Such an almost visceral experience can help inspire new understanding.