Evolution and Structure of the Universe Ripples In Space And Time
Extending simulation of protein on water 100 times longer than prior work offers new insight.


On January 6, 1997, the sun burped, in a manner of speaking, and a billion-ton bubble of hot gas came blasting toward Earth at about a million miles an hour. Four days later, pushing a pileup of solar magnetic field called a magnetic cloud, the bubble rammed the nose of Earth's magnetic field, and while most of us were blissfully unaware, the resulting magnetic storm put on a show for space physicists.

uvi.jpg
This POLAR satellite view of the Arctic shows an ultraviolet image of electromagnetic radiation from the auroral oval during the January 1997 magnetic storm. Image provided courtesy of Marshall Space Flight Center.
Download larger version (277KB) of this image.

A satellite orbiting 350 miles above the Earth recorded an electrical burst 10,000 times normal intensity, and the high levels lasted a week. On January 11, the dense trailing edge of the bubble delivered a parting shot to the magnetosphere — the outer boundary of Earth's magnetism, and the energy dissipation in the auroras over the North and South Poles climbed to 1,400 gigawatts, nearly double U.S. power-generating capacity.

It was an impressive display of space weather, and it demonstrated how these gusts of solar wind — the stream of charged particles from the sun — can play havoc with our information age infrastructure. On January 11, loss of radio contact grounded a plane trying to take off from Halley Bay Station, Anarctica, and a $200 million communications satellite, Telstar 401, fell silent, forcing network broadcasts to switch to another satellite.

Severe magnetic storms have occurred before. A March 1989 storm blacked out Hydro Quebec, a Canadian utility, and the drain on generating capacity threatened to knock out the northeast U.S. power grid. January 1997 was the first time, however, that a fleet of observational satellites was in place to capture data from such an event, and by January 17, less than a week after the solar bubble passed, Chuck Goodrich and colleagues at the University of Maryland Space and Plasma Physics Group (SPP) were using this data in an unprecedented computational study.

"This event had intense scientific interest and news coverage," says Goodrich. "NASA asked us to simulate it, and thanks to PSC we were able to provide video animations on very short notice." Running on four processors of the CRAY C90 at Pittsburgh Supercomputing Center, Goodrich and his SPP colleagues, John Lyon of Dartmouth, Dennis Papadopoulos and Mike Wiltberger used their 3D magneto-hydrodynamics software to simulate the entire storm, from the arrival of the shock front preceding the bubble through the next day and a half, over 45 hours of storm time. "In space science," says Goodrich, "this was a simulation project of unprecedented proportion." Within a week, they had results: seven gigabytes of data, and visualizations that offer a global view of a geomagnetic storm unobtainable any other way.

multi-lanl-0144.jpeg
multi-lanl-0221.jpeg
Earth's Magnetosphere
The top image shows the "relaxed" magnetosphere three hours prior to the onset of the January 10, 1997 magnetic storm, solar wind arriving from the left. The bottom image shows the magnetosphere three hours later. The translucent surface represents the shape and extent of Earth's magnetic field. The colormap indicates plasma density and shows the sunward shock wave around the Earth.
Click on the images to download a larger version.

The Big Picture

First observed in the early 1970s, these eruptions of solar gas occur as often as once a day and the frequency may increase as the year 2000 peak of the 11-year solar cycle approaches. Even the most severe solar gusts present no known health hazard, since Earth's magnetosphere serves as a protective barrier, deflecting the mass flow. But some of the flow energy transmits through the magnetosphere and shows up as bursts of electromagnetism in the ionosphere, hence the Northern and Southern lights, and the potential to seriously disrupt our power supply and communications systems.

The advance guard for detecting blasts of solar wind headed our way was two satellites, SOHO and WIND, that hover about 600,000 miles out on the sunward side of Earth — revolving around the point in space where the Earth's gravity and the sun's are equal. Earth-orbit satellities closer in gather data on near-Earth effects.

While the information from these satellites is invaluable, it represents pinpoints in a huge volume. "You often don't know how what happens here relates to what happens over there," says Goodrich. "You may have readings from several points at the same time, but you don't know what happens in between." With supercomputing, the SPP group connects the dots and produces a big picture view.

"The basic question in magnetospheric physics," says Goodrich, "is how energy from the plasma flow past the Earth penetrates the magnetosphere and gets absorbed in the inner atmosphere. Until recently, we knew observational effects, but we didn't know the basic physics behind them. We have theories, but it's been impossible to check them. Simulations help us understand the detailed physics."

The SPP numerical model simulates a solar storm impacting the magnetosphere and transferring energy through it to Earth's ionosphere. SPP used data from WIND and SOHO for the flow and magnetic field of the January 1997 incoming gust of solar wind. A set of magneto-hydrodynamic equations then takes over and calculates effects throughout the magnetospheric volume, represented as a grid in the stretched-out spherical shape of the magnetosphere. Earth's ionosphere is a lower boundary of the model volume, from which the simulation maps electrical current and velocity through the ionosphere. "From that model," says Goodrich, "we can predict ground currents and magnetometer measurements on Earth, and what's floating onto the electric power lines."

The Unforeseen Effect of Plasma Density

multi-lanl-0144.jpeg
multi-lanl-0144.jpeg
multi-lanl-0221.jpeg
These frames from the SPP visualization show plasma density (indicated by color) in the plane of Earth's equator in the early hours of January 11, when the dense trailing edge of the solar bubble strikes the leading-edge of Earth's magnetosphere. The black circle represents geosynchronous orbit, with satellite positions indicated. Earth is magnified by a factor of two for clarity.
Click on the images to download a larger (200KB) version.

The results of the SPP simulations of the January 1997 storm agree well in key details with Earth-orbit satellite data, confirming basic accuracy of the model. The simulation shows bursts of released energy in the ionosphere happening close in time to when such bursts were recorded at near-Earth satellites. Similarly, the simulation shows energy entering the ionosphere in the same place shown in ultraviolet images from POLAR, a satellite orbiting above the Arctic.

The simulation shows the first significant energy transfer — called a magnetic substorm — occurring on January 10 after a southward turn in orientation of the solar wind magnetic field. Physicists have postulated — and studies have confirmed — that energy transfer happens predominantly when this occurs, since southward orientation of the solar field is opposite to the northward orientation of Earth's magnetosphere. The opposed polarities, in theory, allow the magnetic field lines to connect and energy to flow across the connections.

The simulations confirm this view, and offer added insight. When the magnetic cloud in front of the gas bubble arrived, it brought with it an abrupt southward shift and an increase in the strength of the solar magnetic field, initiating a prolonged substorm. Late on January 10 through early January 11, the field direction turned gradually northward, closing off the energy transfer.

The simulation showed, however, that within the period of southward orientation the rise and fall of ionospheric activity correlated with solar wind density. This was a new finding. "Prior studies of substorm triggering," says Goodrich, "focused on the northward turning of the magnetic field." The simulation shows lack of intense activity for almost two hours after the magnetic cloud arrives. Although it brought a strong southward shift of the magnetic field, the plasma density was unusually low. Subsequent bursts of energy transfer correspond to arrival of denser gas behind the magnetic cloud.

Goodrich and his colleagues surmise that prior studies may not have reflected the extreme southward field orientation that persisted during the January 10 magnetic cloud: "We suspect that the large amount of coupling between the solar wind and the magnetosphere during the southward phase of the cloud allowed a significant fraction of solar wind kinetic energy to enter the magnetosphere." The SPP group plans to investigate this hypothesis in future work.

The January storm was not the biggest of its kind, notes Goodrich, but its significance lies in being the first intense magnetic storm followed in its entirety, from the sun to the ionosphere. "This is the first one we've been able to track as a full event," says Goodrich, "and the animations had a big impact. Space physics deals with invisible fields and particles, and an image carries great weight. Without centers like PSC, we wouldn't be able to do this."

Excerpt from the video animation of CME impact on our magnetosphere (4.2MB). The full animation can be found at the University of Maryland, Space and Plasma Physics Group Website.

Researchers: Charles C. Goodrich, University of Maryland, Space and Plasma Physics Group.
Hardware: CRAY C90
Software: User developed code.
Related Material on the Web:
The Space and Plasma Physics Group at the University of Maryland.
MHD Simulation of January 10-11 1997 Magnetic Cloud Event (includes MPEG animations).
Projects in Scientific Computing, PSC's annual research report.

References, Acknowledgements & Credits


© Pittsburgh Supercomputing Center (PSC)
Revised: October 22, 1998

URL: http://www.psc.edu/science/Goodrich/goodrich.html