Early last month, U.S. officials warned power companies that a minor disturbance on the sun might create slight wobbles in electric currents of power lines.

What followed, instead, was a surprisingly strong jolt in the Earth's magnetic field that forced power operators to quickly tweak the voltage in their systems to avoid blackouts.

The event, as U.S. government space-weather forecaster Joe Kunches recalls, illustrates the limitations and unevenness of current methods for forecasting space storms.

For power grid and satellite operators, "it was like having a tornado sit on top of your house for a day," said Kunches, lead forecaster for the National Oceanic and Atmospheric Administration's Space Environment Center in Boulder, Colorado.

The field of space weather forecasting, increasingly important in an ever more technology-dependent world, is in its infancy. Forecasters can't obtain accuracy anywhere near that of Earth-weather prediction. There are many reasons; but the main one is that space is, well, very big.

Just how difficult is space weather forecasting?

"You've got to realize I'm a forecaster, and I'm probably biased," said Kunches. Even so, he acknowledged, "space weather forecasting is probably something like 50 years behind the maturity level of terrestrial weather forecasting."

But scientists are making headway, and there have been major successes to counter the failures, according to researchers. For instance, Kunches said, his agency gave more than 12 hours' warning of another big geomagnetic storm in mid June.

As the sun hit the peak of its 11-year cycle of storm activity this year, researchers improved their understanding of how solar storms work. And powerful new measuring equipment planned for launch into space in the coming years, they said, will dramatically improve the situation further.

It had better improve, many observers believe, because the past 50 years have seen increasing evidence that solar activity can affect people more than many think.

For instance, according to the Space Environment Center:

• A powerful solar-radiation storm can blast people in an airplane flying at high latitudes with the rough equivalent of as many as 100 chest X-rays. This is likely to happen less than once every 11 years.
• In October 1989, the sun spewed radiation levels high enough to be fatal to a moonwalking astronaut caught in the brunt of the solar storm with only a spacesuit as protection. (Astronauts who had time to seek an underground shelter would have received only a mild radiation dose.)
• Solar weather may affect global circulation patterns and weather, though how is unclear. One known fact is that stratospheric winds near the equator blow in different directions, depending on the time in the solar cycle.
• According to one estimate, better use of space weather forecasts by the power industry could save the United States $365 million per year on average.

There are no definitive statistics on the accuracy of space weather prediction, said Barbara Thompson, a research scientist at NASA's Goddard Space Flight Center in Greenbelt, Maryland who studies solar weather. Various estimates are all over the map, she said.

But she said the main problem has been predicting whether an event on the sun will affect Earth.

"We're still on 'ifs,' not on the 'whens,'" she said. "The 'when' is still pretty sketchy. But it's getting better. The real problem is the 93 million miles (150 million kilometers) between the Earth and the Sun. It's really hard to reconstruct what is happening in between."

There are three major types of solar storms. All three happen most often during the peak of the sun's 11-year cycle. This coincides with an increase in the number of sunspots -- relatively dark, cool spots on the solar disk.

One type of event is a solar-radiation storm. This is a sharp increase in the amount of protons -- a basic subatomic particle -- in the solar wind, which is a fast stream of gases ejected by the sun. Radiation storms can maim astronauts and damage satellites.

Another is a geomagnetic storm. This occurs when the sun shoots bubbles of hot, magnetized gas at us, jolting Earth's magnetic field and wreaking havoc with power grids and Earth-satellite communications. These events also cause the spectacular aurorae, also known as the northern and southern lights.

The third phenomenon is a solar flare, in which the sun spurts X-ray blasts that can arrive at Earth at light speed and knock out ground-to-air and ground-to-ship radio transmissions.

The current equipment for forecasting space weather consists mostly of two satellites that are gravitationally held at a place in the line between the sun and Earth, though much closer to Earth. They are the Solar and Heliospheric Observatory (SOHO) and the Advanced Composition Explorer (ACE).

Their basic problem is that although they can measure very well what's going on directly around them, that doesn't tell us much until the storm is almost here, explained Richard Canfield, a research professor at Montana State University at Bozeman.

Also, "they make images which show coronal mass ejections" -- spurts of solar gas that foster magnetic and possibly radiation storms, Canfield said. "But they can't even tell simple things like, is the ejection headed toward us or away from us."

The problem may be largely solved by a set of twin spacecraft that NASA plans to launch in 2004. These would view the sun from two points along the Earth's orbit, like a pair of eyes. These would produce vastly improved, three-dimensional images.

Other improvements in forecasting have already come from information gleaned from the current solar cycle peak. Researchers from Catholic University of America and NASA's Goddard Space Flight Center announced in June that they had come up with a precise way to estimate how long it would take a solar eruption's effects to reach Earth, based on the solar-wind speed.

With this type of improvement in modeling techniques, combined with the STEREO mission, "hopefully, any ambiguity will be resolved," Thompson said.