The capacity factor of wind plants in the United States has improved significantly in recent years, with newer plants exhibiting higher capacity factors than older plants. Some researchers have attributed this increase to technology improvements that boost energy capture per unit capacity, while others have instead argued that the role of technology may be overstated, and that a substantial portion of the gains in wind plant output are instead due to a global trend of increasing wind speeds.
A new study from Berkeley Lab, appearing in the journal Wind Energy, utilizes an innovative pairing of surface wind observations with wind plant generation records to investigate these issues.
It is true that ground-level wind speeds appear to have increased in recent years, according to widely available weather station data, measured at 10 meters off the ground. But do these ground-level trends imply that benefits of wind plant technological innovation have been overstated? More generally, do ground-level wind speed trends even accurately represent trends at the elevated heights at which wind turbines operate, around 100 meters above the ground? Since 100 meter wind speed data is not widely available, these are challenging questions to answer.
We explore each question in turn. First, we simply assume that the long-term trends in 100-meter wind speeds match the trends measured across surface weather stations near U.S. wind plants. This assumption allows us to compare possible performance gains from increasing wind speeds to performance gains from technological improvements. Second, we pair individual plant generation records to near-by ground-level weather stations to assess, in general, how closely surface wind speed trends match long-term trends in recorded generation.
Part 1. Performance improvements from technology versus increasing wind speeds
Even assuming 100-meter wind speed trends exactly follow ground-level trends, increasing wind speeds provided only a small boost (2.6%) to the average output of wind plants from 2010 to 2019. This is reflected in Figure 1, which shows only a mild upward trend in surface wind speeds from 2010-2019.
Figure 1. Average surface wind speeds trends (A: actual average wind speeds, B: wind speeds normalized by their long-term average) measured at the network of publicly available surface stations (height ~10 meters). Stations were grouped by distance from a wind plant. Stations closer to wind plants averaged higher speeds compared to more distant stations, but, interestingly, each group had similar long-term trends in wind speeds.
Figure 2. Newer vintages of wind plants (orange, yellow, red) have capacity factors that are much higher than older vintages (blues and gray). Note, to avoid staged construction and other ‘teething’ issues, capacity factors are calculated beginning in each plant’s second full calendar year of life.
As a result, our study concludes that the vast majority of capacity factor improvements over this period are attributable to technology, plant design, and site selection, rather than to changes in underlying wind speed.
Part 2. Do surface wind speed trends accurately represent 100-meter wind speed trends?
The study also digs in to the question of how surface wind speed trends relate to wind generation trends. Specifically, the study looks at whether nearby ground-level wind speed measurements (i.e., measurements within 25 km of a wind plant) provide a good proxy for year-to-year change in recorded wind generation. It finds that year-to-year changes in surface measurements accounted for less than 20% of the year-to-year changes in recorded generation at nearby plants. The implication of this finding is that the publicly available network of surface meteorological observation stations should not be exclusively relied upon to develop a proxy for long-term trends in wind generation.
Why is that? First, winds at “hub height,” or 100 meters off the ground, act differently than surface winds. For example, hub-height winds can be affected by low-level jets, which are fast-moving layers of air that are elevated above surface measurement devices but that are low enough to interact with wind turbines. Another example, changes to surface properties, such as land clearing or building construction, can have a bigger impact on surface wind speeds than hub-height wind speeds. Second, the public network of surface measurement stations is not co-located with wind plants. Wind plants are often sited to take advantage of very specific geographic features that result in higher wind speeds, but weather stations are sited with different criteria – the difference in siting criteria could cause differences in long-term trends. And finally, recorded wind generation is likely to show some variation due to random differences in annual maintenance needs.
Another way to assess annual variation in a wind resource is to use meteorological or reanalysis models to gain insight into hub-height wind speeds. Though this study did not examine these models, prior research has shown that these models do not always agree with one another. In a number of cases, the models do not recreate the annual wind speed trends observed at the surface. The conclusion here is that, at the moment, there is no perfect approach to measuring long-term trends in hub-height wind speeds. Enhanced understanding of long-term trends in hub-height wind speeds could be gained through greater sharing of data by wind plant operators or owners, or widespread measurement and modeling campaigns—both likely to be challenging. In the meantime, smaller research efforts can continue to provide additional insight on this topic.
Article and Contact Information
The article in Wind Energy, “What Can Surface Wind Observations Tell Us About Interannual Variation in Wind Energy Output?” is ‘open access’ and available to all: https://doi.org/10.1002/we.
If you have any questions, please contact Dev Millstein, DMillstein@lbl.gov, at Lawrence Berkeley National Laboratory.
We appreciate the funding support of the U.S. Department of Energy’s Office of Energy Efficiency and Renewable Energy.