Scaling of wind energy variability over space and time

We use a large data set of simulated wind energy production in the United States to quantify the geographic and temporal scaling of energy output variability reduction when multiple sites are aggregated. We add to the existing literature on “geographic smoothing” by (i) quantifying the scaling of geographic smoothing over multiple spatial and temporal scales; (ii) bounding such smoothing through the use of an algorithm that produces minimum-variance sets of wind energy production sites; and (iii) quantifying inherent tradeoffs in optimizing wind energy site selection to minimize output variability along a specific frequency. The number of wind farms required to minimize output variability increases linearly with spatial scale of aggregation, but the scaling factor is small, on the order of 10-6 relative to geographic distances. These scaling factors increase by a factor of two as the frequency considered increases by three orders of magnitude (minutes to months). Our analysis indicates that optimizing wind deployment over one particular frequency increases output variability over other frequencies by nearly 30% in some cases.