July 17, 2003
Squeezing more out of renewable energy
By PASCAL STORCK
3TIER Environmental Forecast Group
Energy-related headlines are increasingly hard to ignore these days.
The Federal Reserve Chairman is worried about the high price of natural gas and the potential dampening effect on America’s economy. Our reliance on imported oil complicates our political interests around the world. Our overwhelming dependence on fossil fuels clouds the debate on global warming and makes concrete action economically difficult.
In the face of these concerns, renewable energy is often touted as the solution and nowhere in our country is this debate more relevant than in the Pacific Northwest. While the percent of total electricity generated from all renewable sources in the entire U.S. is about 9 percent, in the Northwest we rely on renewable sources, mainly hydropower, to generate 71 percent of our power.
Renewable energy concerns
Renewable energy sources are not without their concerns. Since carbon-based fuels like coal, oil and natural gas can be easily stored and purchased on demand, electricity can be generated from these sources at will with the only immediate concern being the price of the fuel source. If a coal-fired power plant is running low on fuel, the operators can always order more.
Herein lies one of the great challenges with renewable energy sources. A hydropower operator can’t simply order more water if the reservoir runs low. A wind farm operator can’t request more wind if they’d like to generate power at full capacity tomorrow.
While these sources of renewable energy liberate us from our dependence on fossil fuels, they make us dependent on something entirely different: the weather. As we increase our dependence on renewable energy we need to be able to predict how much generation capacity we will have in the future, not only for the next hour or day or month, but for the next year and decades into the future.
In short, we need to understand and predict the weather and climate.
Beyond the weather
Northwestern natives know how difficult our weather is to predict. Given the dependence of renewable energy on weather, this difficulty has lead to the unfair characterization of renewable energy as an unpredictable resource.
Even though a forecast of cloudy with a chance of showers and sun breaks seems to do just fine on most days for most people, utilities require more accurate information on exactly how the weather will affect their generation resources.
For example, a wind farm operator is not really concerned with how windy it will be, but instead needs to know how much power they will produce. Similarly, a hydropower operator is not really concerned with how much it will rain next June, but instead needs to know how much water will flow into their reservoir.
Both wind power production and reservoir inflow depend on a myriad of factors beyond the weather. Also, since most renewable energy resources are located away from major metropolitan areas, these utilities have generally been unable to obtain forecasts for their specific project.
To address these needs, 3TIER Environmental Forecast Group develops, maintains and operates a comprehensive environmental forecast system to provide forecast information for its clients. We have found that recent advances in computer power, coupled with advances in environmental simulation models and improvements to global and local observational systems allow us to predict the weather and the state of the environment accurately for a specific site and a specific resource.
Examples of forecast techniques are presented below for the Pacific Northwest’s two most important renewable energy resources, hydropower and wind power.
Hydropower. The grandfather of all renewable energy resources in the Pacific Northwest is hydropower. The forecasting of reservoir inflows has a long history with all forecast techniques relying on some combination of snow survey data and/or climate indicators.
Forecasts based on snow survey data attempt to match current snow data to historic snow packs and their historic streamflow during the summer months. While these techniques are accurate late in the snow year, they are not very useful early in the water year (October or November) or outside the summer runoff period. For example, it is nearly impossible to make a forecast based on the snow pack before any snow has fallen.
Climate signal forecasts draw from our experience that the climate of the Pacific Northwest is connected to the sea surface temperatures of the equatorial Pacific Ocean. During years categorized as “La Nina,” the Northwest generally experiences higher than average winter snow pack and above average summer streamflow. The opposite is true during “El Nino” years.
The advantage of climate forecasts is that a useful forecast can be made in advance of the snow accumulation season.
Today’s state-of-the-art streamflow forecast models integrate current observations from the watershed (the snow survey technique) with the long-term weather forecast (the climate signal) and short-term weather forecast information. By integrating these techniques, each source of information can be used to its fullest potential, which results in a more accurate streamflow forecast. For example, during the winter of 2001/2002, 3TIER’s forecasts for Seattle City Light suggested that summer runoff in the Pacific Northwest would be much higher than average as early as mid-December 2001 when other forecasts were calling for below average runoff.
Wind power. Forecasting for wind energy is quite different than for hydropower. While most hydropower projects have considerable storage capacity and therefore have some control over when they generate power, wind can’t be stored and the energy generated must be sold and utilized the instant it is created. Therefore a wind power operator is much more concerned with short-term environmental forecasts.
Forecasts for the next hour to several hours are useful for satisfying generation commitments, while forecasts for the next day to several days are useful for transmission scheduling and obtaining the best price for generated power.
To support its wind energy clients, 3TIER maintains and operates several high-resolution numerical weather prediction models configured for a specific wind client’s project. The forecasts from these models are then used as input to an adaptive statistical model that predicts wind energy production at the site.
When compared to simple forecast strategies, these techniques double the accuracy with which wind power can be scheduled at lead times as far as several days in advance. A few hours in advance, these strategies can improve the scheduling of wind by 10 percent or more.
While these improvements may seem modest, a 10 percent increase in next hour forecast accuracy is roughly equivalent to a savings of $1 million per year for a 100-megawatt wind farm.
Perfect fit or perfect storm?
As our reliance on renewable energy increases in the Pacific Northwest, a major concern becomes how renewable energy resources behave together.
Considering wind and hydro power, one can imagine the perfect fit scenario. During those years in which the Pacific Northwest is having a major drought, our wind farms produce at full capacity and offset the lack of hydropower generation.
However, one can also as easily imagine the perfect storm. During a summer drought, the wind refuses to blow and our wind farms sit idle. During the perfect storm scenario, we can’t rely on the wind energy resource to balance the hydro resource.
Scientists and researchers are beginning to address the seasonal correlations between wind and hydropower and the initial results point to a strong synergistic relationship between wind and hydropower.
Nevertheless, the dependency of renewable energy on the weather and climate underscores the importance of environmental forecasting in renewable energy’s future.
Copyright ©2009 Seattle Daily Journal and DJC.COM.
Comments? Questions? Contact us.