“It's cloud illusions I recall,” Joni Mitchell sang in 1969. “I really don't know clouds at all.”

A sentiment undoubtedly shared by environmental engineering professor Jan Kleissl, who says knowing where clouds will cluster in the next three hours is becoming increasingly important as solar energy forms a larger part of the nation’s electricity generation mix.

Kleissl, of the University of California at San Diego (UCSD), is currently using his sky-tracker computer model, satellite images and algorithms to see if he can predict the movement of clouds in the very short term.

This is important because distributed solar, on rooftops for example, is superb at keeping individual neighborhoods powered up when the sun is visible, but becomes a liability for both users and power companies when clouds block the sunlight.

When that happens, utilities are forced to switch on expensive, natural gas “peaking” generators to cover the gap, since the larger power plants are already working at capacity. And this – the extra expense and the unmanageability of intermittent power sources like solar and wind – is one of the main reasons why electric utilities oppose renewable energy. It is unpredictable, and difficult and costly to store.

Kleissl’s project seems like an extremely ambitious undertaking, but perhaps not insurmountable thanks to $550,000 from California electricity ratepayers and another $137,000 from unnamed sources, not to mention the assistance of regional utility San Diego Gas & Electric Co., or SDG&E; the California Independent System Operator, or CalISO, which oversees the state’s power grid; and EDSA, a San Diego-based power analytics solutions provider.

As the basis for his model, Kleissl takes pictures of the sky surrounding the university every 30 seconds. He also uses satellite images and information provided by an on-campus solar array. Later, when all the data is assembled and fed into the program, the algorithms may offer some clue as to the behavior and timing of cloud formations. If so, it would be the first time in history, and will, according to CalISO policy analyst Jim Blatchford, encourage even more solar power.

If not, energy experts believe that the power grid can operate in stable fashion as long as 85 percent of the electricity is derived from conventional resources; that is, utility-scale, fossil-fuel burning power plants.