Researchers and Xcel Energy have joined forces to try to tighten up one of the key uncertainties in the renewable energy business: the weather forecast.

Using tools that could provide forecasts for wind, icing and other dynamics down to the location of individual wind turbines, the National Center for Atmospheric Research (NCAR) and Xcel are aiming to make better predictions about how much power the utility's turbines might produce and when. They're also investigating better ways to measure solar radiation, so they can calculate more exactly how much power they don't need to provide at times to homes and businesses with their own solar-energy systems.

Both could help Xcel, the nation's top wind power utility, save millions of dollars, said Drake Bartlett, senior trading analyst for Xcel. More accurate and more up-to-the-minute wind and solar forecasts could help Xcel better plan its needs for backup generation units fired by coal and natural gas.

National Weather Service forecasts apply to areas about 3 miles square. And wind speeds are predicted in hourly increments "at the human being level," said Dan Luna, meteorologist in charge at the Weather Service's Twin Cities office. But that doesn't meet a utility's needs, noted Melinda Marquis, renewable energy program manager for NOAA's Earth Systems Research lab, which, like federally funded NCAR, is based in Boulder, Colo. That's true for many industries, which hire private forecasters and other analysts to get a better sense of the weather's potential effects on their bottom lines, she added.

The Xcel-NCAR project will instead develop continuous forecasts for wind speeds, temperature and other conditions at the nacelle of the wind turbine, the box containing the generating equipment, where the turbine's blades meet. Nacelles are situated about 80 meters above the ground, where the weather can be quite different than it is below.

The two partners have been exploring that type of fine-scale forecasting since 2009. But over the next two years they will be tightening the focus even more, looking for ways to forecast "ramps," or sudden increases in winds; icing, using NCAR experience in aviation; and solar radiation, which will involve tracking clouds and cloud types, which all have varying effects on how much of the sun's energy is available for use.

Xcel Energy officials say the more accurate forecasts are key to the company's increasing use of renewable energy sources. Because large-scale electrical energy from any source can't be cost-effectively stored, producers have to continually balance production with demand. But if Xcel comes up short on its projected contributions from renewables to the regional electrical grid, the company has to pay other producers to make up the difference, Bartlett added. So Xcel needs to know more precisely how much power it might get from wind and solar sources.

Xcel is providing NCAR with data that it owns from sensors on its wind turbines, and NCAR will use that in developing and enhancing weather forecasting models for the utility. Data-sharing between corporations and the NOAA (which includes the weather service) is becoming increasingly common, Marquis added, even with proprietary business information.

"There's plenty of room for the private sector to take weather forecast output and improve it," Marquis said.

But because NCAR will ultimately share its research with the scientific community, the project with Xcel — particularly the research into cloud forecasting — could lead to improvements in weather forecasts for the general public "beyond just for renewable energy," said Sue Ellen Haupt, weather systems and programs assessment director for NCAR.

Staff writer Dave Shaffer contributed to this report.

Bill McAuliffe • 612-673-7646