ML102930306

Exhibit 8 of 21-Electric Power from Offshore Wind Via Synoptic-Scale Interconnection

ML102930306
Person / Time
Site:	Seabrook
Issue date:	02/23/2010
From:	Colle B, Kempton W, Pimenta F, Veron D Univ of Delaware
To:	NRC/SECY
SECY RAS
Shared Package
ML102930267, ML102930301	List: ML102930302 ML102930303 ML102930304 ML102930305 ML102930306 ML102930310
References
50-443-LR, RAS 18937
Download: ML102930306 (6)

Text

Electric power from offshore wind via synoptic-scale interconnection Willett Kemptona,1, Felipe M. Pimentaa, Dana E. Verona, and Brian A. Colleb a

Center for Carbon-free Power Integration, College of Earth, Ocean and Environment, University of Delaware, Newark, DE 19716; and bSchool of Marine and Atmospheric Sciences, Stony Brook University, Stony Brook, NY 11794-5000 Edited by M. Granger Morgan, Carnegie Mellon University, Pittsburgh, PA, and approved February 23, 2010 (received for review August 14, 2009)

World wind power resources are abundant, but their utilization power fluctuations are important if wind is to displace significant could be limited because wind fluctuates rather than providing amounts of carbon-emitting energy sources.

steady power. We hypothesize that wind power output could There are four near-term ways to level wind power and other be stabilized if wind generators were located in a meteorologically fluctuating generation sources. (i) Expand the use of existing con-designed configuration and electrically connected. Based on 5 yr of trol mechanisms already set up to handle fluctuating load and wind data from 11 meteorological stations, distributed over a unexpected equipment outagesmechanisms such as reserve 2,500 km extent along the U.S. East Coast, power output for each generators, redundant power line routes, and ancillary service hour at each site is calculated. Each individual wind power genera- markets. This is how wind is integrated today (5). (ii) Build energy tion site exhibits the expected power ups and downs. But when we storage, as part of the wind facility or in another central location.

simulate a power line connecting them, called here the Atlantic (iii) Make use of distributed storage in loads, for example home Transmission Grid, the output from the entire set of generators heaters with thermal mass added or plug-in cars that can charge rarely reaches either low or full power, and power changes slowly. when the wind blows or even discharge to the grid during wind SUSTAINABILITY Notably, during the 5-yr study period, the amount of power shifted lulls (6). (iv) Combine remote wind farms via electrical transmis-up and down but never stopped. This finding is explained by ex- sion, the subject of this article.

Prior Studies of Wind Leveling via Transmission. Several SCIENCE amining in detail the high and low output periods, using reanalysis data to show the weather phenomena responsible for steady pro- studies in the western United States and Europe have investi-duction and for the occasional periods of low power. We conclude gated the power leveling of aggregating geographically distribu-with suggested institutions appropriate to create and manage the ted wind farms (7-10). They find improvement in the steadiness power system analyzed here. of the available power, even when the stations are relatively close (11, 12) and a decrease in the number of low- or no-wind events meteorology transmission wind integration wind power (13, 14). Additional stations decrease the variability of the meteorologically designed transmission summed wind power (8). Reduced variability means fewer very low power times, as well as fewer times of the highest power (14).

Interconnecting wind generators generally yields greater T he worlds wind resource for electric power is larger than the total energy need of humanity. For surface winds over land globally, Archer and Jacobson (1) estimate the wind resource benefit for longer separation distances. There is less benefit from proximate stations, as they are more likely to experience similar at 72 terawatt (TW), nearly five times the 13 TW worlds demand weather at the same time, due to local forcing conditions such as for all energy. In a more detailed regional estimate, Kempton et changes in topography or surface high- and low-pressure systems.

al. (2) calculated that two-thirds of the offshore wind power off Greater distances between wind stations usually lead to longer the U.S. Northeast is sufficient to provide all electricity, all light- periods of smoothing (15, 16). For example, local geographic vehicle transportation fuel, and all building heat for the adjacent dispersion in Germany has been shown to smooth on short time-scales (5 min) at station distances of 2 km. Some studies (8, 10, states from Massachusetts to North Carolina.*

Planning of wind development in the U.S. Atlantic region is 13) suggest there will be a distance, roughly 800-1,000 km, beyond already underway. Fig. 1 shows as black squares offshore wind which adding a station no longer brings additional improvement.

For example, Oswald et al. (10) analyzed eight stations distri-developments that have already been approved by their adjacent buted throughout Britain during 12 Januaries, a month of peak state governments. Each square represents a planned array of power demands, peak wind speeds, and peak variability. They 80-150 turbines, with each array having a capacity of 280-also compared Britain with data from Ireland, Germany, and 425 megawatts (MW). Together these represent a power capacity Spain. Observing large power swings in a 12-hour period, Oswald of about 1,700 MW (the scale of a large coal or nuclear power et al suggested that distributed generation would not help much plant), yet together they tap only 0.1% of the regions offshore since most of the region experienced the same wind conditions.

wind resource (2). Each will be connected independently to But, one might ask whether their grid orientation and size were the electric grid by a submerged power transmission cable running ashore to the closest transmission. Electric system plan-ning for each has proceeded separately, to meet the power needs Author contributions: W.K. designed research; W.K., F.M.P., and B.A.C. contributed new reagents/analytic tools; W.K., F.M.P., D.E.V., and B.A.C. analyzed data; and W.K., F.M.P.,

of each adjacent state. Here we analyze the spatial and meteor-D.E.V., and B.A.C. wrote the paper.

ological aspects of distributed offshore generation, then conclude The authors declare no conflict of interest.

by proposing a more coordinated regulatory approach better This article is a PNAS Direct Submission.

matched to this power resource.

Freely available online through the PNAS open access option.

Leveling Wind Fluctuations. The variability of wind power is not as problematic as is often supposed, since the electric power *If wind power were deployed at the scale implied by these resource studies, it would affect weather (3) and climate (4). Nevertheless, the global effects of even very large system is set up to adjust to fluctuating loads and unexpected wind power deployment appear to be more modest and more manageable than the failures of generation or transmission. However, as wind power effects of climate change (4).

becomes a higher proportion of all generation, it will become 1 To whom correspondence should be addressed. E-mail: willett@UDel.edu.

more difficult for electric system operators to effectively integrate This article contains supporting information online at www.pnas.org/cgi/content/full/

additional fluctuating power output. Thus, solutions that reduce 0909075107/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.0909075107 PNAS Early Edition 1 of 6

simplifying the administrative possibilities for planning transmission interconnection and coordinated development. The methods for computing power output from wind speed are found in Materials and Methods below. Since the wind speed data we use are typically reported with 1 h resolution, we cannot evaluate the effect of smooth-ing and the continuity of supply over much shorter time intervals.

Results We first analyze the seasonal patterns of wind energy and the cor-relation of power from different stations distributed along the coast. Next we explore the effects that offshore transmission could have upon aggregate power produced by the entire array.

Calculating Power Generation and Capacity Factor. Summary averages over the entire 5-yr study period are shown in SI Text, Table S2, specifically, the average wind speed, the average power output for an example 5 MW turbine (Fig. S2), and the capacity factor, or CF. The capacity factor is a standard measure for wind power analysis. It is a dimensionless quantity defined as energy out-put per year (in MWh) divided by the number of hours in the year times the nameplate, or rated, capacity (in MW). In simpler terms, CF is the average output (in MWa ) over the rated capacity (in MW)

(14, 22). For example, if a 5 MW turbine produces an average of 2 MWa throughout a year, its CF in that location is 0.40. These methods are described in prior publications (20) and in SI Text.

Some of our figures below will use stations S2 and S10 as dis-tinct illustrative examples. Station S2 is the second lowest in CF and average power output, whereas S10 is the second highest.

They coincidentally are also each one away from the ends of the extent, thus experiencing different weather. Thus we use stations S2 and S10 in some subsequent figures as examples of Fig. 1. Proposed offshore wind projects off the U.S. coast (black squares and names). NDBC offshore meteorological stations selected for this study, diverse individual sites. For summary and quantitative measures, running from Florida (S1) to Maine (S11) with color bullets for buoys and col- we continue to use all 11 stations.

or triangles for towers. Inset photos: discus buoy (S8), with anemometer height of zref 1/4 5 m, and a lighthouse tower (S2) at zref 1/4 44 m. Correlation of Wind Power Output with Distance. Previous studies have shown that wind speed correlation between stations drops sufficient for this conclusion. The synoptic pressure patterns of off with distance (7, 10, 13). Here we compute the Pearson Europe (17) and the United States (18) are larger than the extent correlation in electric power output among stations.

of Britain as a grid studied by Oswald et al (10). In Fig. 2, each dot is a pair of stations, with each pair plotted by Perhaps these studies were not of sufficient extent or meteo- distance between them and correlation between their wind rological diversitythis is an important question for the analysis speeds. The highest correlations, r > 0.6, occur for stations less in this article. For example, during the warm season synoptic than 350 km apart (the dot at 0 distance and 1.0 correlation re-high-pressure areas with light and variable winds can extend over presents all 11 stations, each correlated with itself). For stations 1; 000 km, thus we hypothesize that a distributed grid must be more than 750 km apart, correlations are below 0.2, and more

>1; 000 km in order to achieve nearly continuous power. Our than 1,300 km, correlations are below 0.1. This confirms our first second hypothesis is that the orientation of the grid is important hypothesis, that correlation drops at the synoptic scale.

in order to maximize the diversity in the regional meteorology. For mitigating wind power fluctuations, the ideal would be to For example, since many U.S. East Coast cyclones frequently combine stations with negative correlation coefficients, as sug-track along the coast from southwest to northeast, we will here gested by Kahn (7). For negatively correlated pairs, when one is test whether a transmission grid along the coast achieves more high, the other is likely to be low, yielding more steady combined smoothing than other orientations in this region.

Some studies propose methods for selecting among stations to make an optimum aggregate (7, 19, 20). Here we address the more fundamental question of variability and combination of stations at the synoptic scale and leave optimization of station selection for subsequent analysis. [We use the term synoptic to refer to wind fluctuations that are due to the passage of large-scale (1; 000 km) high- and low-pressure systems.]

Meteorological Choice of Region. Our approach is distinct from prior work in that we examine a region larger than the synoptic scale (Fig. 1), not subject to uniform meteorological conditions (Fig. S1),

and more aligned along the prevailing movement of high- and low-pressure systems rather than perpendicular to it. All the stations in our study area (Table S1) are over the ocean, which has stronger and more constant winds than land (21-23). An additional practical rea-son for the choice of this continental shelf to study is that it is adjacent to one of the most carbon-intensive urban concentrations in the Fig. 2. Correlation R between pairs of stations, by distance. The gray curve re-world, yet these populations comprise a single national entity, presents an exponential fit R 1/4 expxD for the optimum value of D 1/4 430 km.

2 of 6 www.pnas.org/cgi/doi/10.1073/pnas.0909075107 Kempton et al.

power. We find only one negative correlation within the hourly Power from the Atlantic Transmission Grid: Statistics Over Five Years.

data (S4 with S10), and that one is of such small magnitude We statistically summarize the effects of transmission intercon-(0.001) as to be effectively zero. [Sindens UK study similarly nection via boxplots and histograms, each based on 5 yr of data.

found only one negative correlation, between stations at 900 km Fig. 4 shows the boxplot distributions of CF for two sample sta-separation (13).] Therefore, near-zero correlation may be a more tions (S2 and S10) and for the entire grid. Looking across the realistic goal than negative correlation for pairing of stations. 12 months, Fig. 4 illustrates that S2 and S10 differ in monthly The above results suggest that the intermittency of wind power output, although both show lower power in the warmer months generation might be smoothed and leveled by combining the (May through September). The main point of Fig. 4 is that the output of different geographical stations at distances more than bottommost plot, the entire grid, shows a much smaller interquar-750-1,300 km. Next we test this hypothesis. tile range, indicating that the grids power output frequently falls near the median.

An Offshore Grid to Level Power Generation. The power from off- In Fig. 5, the distribution of output is shown with three histo-shore wind generators can be interconnected by submerged grams of all hours for all 5 yr. The top two histograms show power high-voltage transmission cable. For such long distances (S1 to output from two sample sites, S2 and S10; the bottommost shows S11 would be 2; 500 km), undersea high-voltage direct current output from the entire Grid. The individual sites are choppy, each (HVDC) cables are well-suited. Example prior uses include the with an extreme modal value (0 or 1). The Grid has a modal 500 kV, 3,000 MW Neptune cable connecting New Jersey to Long output value about equal to the average CF, that is, power output Island (24) and NorNed, Connecting Norway to the Netherlands, is most often in the range of midvalues. This is illustrated and the longest submarine power cable to date at 580 km. In the discussed further in Fig. S3, using a probability distribution European Union, undersea cables connecting offshore wind function like that used in some prior studies (2, 14).

farms and multiple countries are under discussion (called the We quantitatively examined the power from the Atlantic Super Grid), but we find no published meteorological analysis Transmission Grid year-by-year over the entire 5 yr, 1998 to of the smoothing effects. We refer to our hypothesized long- 2002 (shown in Fig. S4). During each of years one through five, distance transmission cable here as the Atlantic Transmission CF was below 0.05 for the following percentages of hours: 2.7%,

Grid and the power output from it as Pgrid . 0.6%, 1.3%, 0.7%, 0.3%, or for the entire period, overall, under SUSTAINABILITY As a graphical illustration of how such interconnection affects 0.05 CF an average of 1.1%.

SCIENCE wind power fluctuations, Fig. 3 shows an example month, Novem- During the entire 5-yr time studied, Grid power never drops to ber 1999, and just two of the individual stations to simplify this fig- zero, that is, power output is uninterrupted. Although a zero-ure. The top half of Fig. 3 shows power output from individual output hour may be found if more years are examined, 5 yr of stations S2 and S10 (thin colored lines), compared with power out- interrupted power has not been seen in the prior wind trans-put from the entire grid of all 11 stations, Pgrid , the thick black line. mission analysis reviewed above, nor for that matter is it seen The two individual stations exhibit frequent changes in power out- in individual fossil power plants, which average a 5.6% forced put, even from zero to full power, or vice versa, in a few hours. The outage rate (5). In the next section we will analyze the meteo-bottom half of Fig. 3 shows the change in power output, with each rology and the dynamical reasons for good and the poor periods.

vertical line representing 1 h. For example, a line up to 0.5 means that the output increased by 50% of capacity within 1 h. For indi- Examination of Weather Patterns. To explore the synoptic variability vidual stations S2 and S10, about 20 times in this month the output of winds along the eastern seaboard, the daily North American changed by over 50% in 1 h. By contrast, the entire grid, the black Regional Reanalysis (25) was used, which provides long-term line, changes by no more than 10% of its capacity in any 1 h. winds and pressures for the North American domain at 32-km Fig. 3 shows that in this example month, the Grid would have grid spacing.

improved the generation of electricity by offshore winds in two We analyze sea level pressure and surface winds to track the distinct ways. First, output fluctuates more slowly. The impor- evolution of synoptic systems over the area and to understand tance to the power sector of slower output fluctuations is that their impact on power generation. We select two months, cover-other generators or transmission can be ramped up or down with ing relatively low wind (May) and high wind (November) condi-plenty of lead time. This makes the power from the wind aggre- tions, and a third month to exemplify an unusual period of gate more valuable and easier to manage than the power from prolonged very low energy generation (June 1998). All are picked wind at a single location. The second improvement is that the also because they have no gaps in the National Data Buoy Center Grid produces midlevel power more often than extremes, an data (NDBC) time series, so with each we can compare a months effect also noted in prior studies. generation of power across all stations.

Fig. 3 illustrates these important principles but covers only one In order to understand these patterns of variability, we now turn month. In the next section we extend this analysis to a 5-yr period, to the daily reanalysis data. We pick four illustrative meteo-using aggregate statistics. rological events, two on Fig. 6 and two on Fig. 7. In 6A and 7A, each Fig. 3. (Top) One month of power, ex-pressed as CF, from two isolated wind parks (blue and or-ange lines) compared with power from the Atlantic Transmission Grid (P grid , thick black line). (Bottom) Hourly changes in CF, compa-ring individual wind parks (blue and or-ange) with the Grid (black line).

Kempton et al. PNAS Early Edition 3 of 6

Fig. 5. CF histograms for individual locations S2, S10, and P grid .

southern stations S1-S3. Meanwhile, weaker winds are observed at the center of the high-pressure positioned over North Carolina.

While moving eastward, this high-pressure still generates full rated power (CF 1/4 1) for nearly a week for the northern (S8-S11) and southern (S1-S3) stations. This example also illustrates why it is Fig. 4. Capacity factor monthly variability for individual wind parks at S2, important for the N-S Grid to extend a distance longer than the S10, and for P grid, shown as boxplots. The central mark in each box is the med- synoptic scale of the Bermuda High.

ian; box edges are the 25th and 75th percentiles. The whiskers extend to the Two examples of less ideal power leveling are shown for June most extreme data-points not considered outliers (approximately
2.7 and 1998 on Fig. 7. During June 13-16, a wide low-pressure system 99.3% coverage if data is normally distributed). Outliers are plotted as circles.

develops over the eastern United States (Fig. 7B). This is the result of a smaller cyclone over the northeast United States on event is shown by a gray band over the power curves. In 6B and 7B, June 13-14, and then another cyclone approaching from the each event is expanded, using NARR data, to show high and low Great Lakes that merges with this northeast low-pressure system pressure with H and L, the sea level pressure with lines for hPa, on June 15. As a result, the stations to the north (S10-S11) and and 10-m wind speeds with color for m s1 , as shown in the scale. center of the grid (S4-S6) experience reasonable generation of For the first gray dates in Fig. 6A, May 1-4, 1999, the NARR power, others do not (Fig. 7A, first gray band of dates). Yet data in 6B show that an extratropical cyclone moved northward the grid output is fairly level except midday on June 15.

along the East Coast. Strong northeasterly winds (14-17 m s1 ) Aworse situation for wind generation is June 21-27, 1998, when a were off the North Carolina coast on May 1, 1999. As the system broad area of weak high pressure (1018 hPa) develops over the mid-moved northeastward on May 2-4, the entire set of northern Atlantic by June 25 and remains steady above the region for nearly stations (S5-S11) experienced winds of 8-12 m s1 . Turbine 4 d, June 25-27 (Fig. 7B). Because of the large dimensions and very power during May 1-4 is seen in Fig. 6A. Stations S4-S6 are weak pressure gradients of the high-pressure system, wind speeds at full power on the first of May, a day when northern stations are generally less than 5 m s1 throughout the entire area.

have very low production (compare CF of each station during this To better understand the synoptic flow associated with low-event). As the system moves northward, on May 3, S4 power falls power periods, we selected intervals of less than 0.05 CF and com-quickly, followed by S5 and S6. Meanwhile, as the weather system posited (averaged) those. The resulting composite pressure maps enters the Gulf of Maine, stations S9, S10, and then S11 start were too smooth to make any physical interpretation; that is, they producing energy around May 2, sustaining it for 2-3 d. did not show a clear pattern of a predominant type of synoptic Fig. 6 also shows the synoptic situations for November 4-9, 1999. situation when low winds occur throughout. After perusal of A large anticyclone remains relatively stationary over the East several of these low-power events, we infer that widespread Coast from November 4-9 (Fig. 6B). The anticyclone over the low winds could be due to any of several causes: an extension southeast intensifies from 1027 to 1031 hPa between November of the Bermuda high, or a separate high moving to the north 4-6, 1999. Subsequently, the high-pressure system weakens to or south, or just a baggy low over the East Coast.

1025 hPa by November 9, 1999. During this period, higher winds Overall, the daily weather data illustrate why the Atlantic Trans-are found on the northern and southern sides of this system, respec- mission Grid yields uninterrupted power output. There is almost tively, over coastal New England and Florida, where the largest sur- always a pressure gradient somewhere, and cyclonic events move face pressure gradients are found. This generates relatively strong along the coast. There are a few times of low power throughout, eastward winds for stations S10-S11 and westward winds for the but they are not due to any one particular weather pattern.

4 of 6 www.pnas.org/cgi/doi/10.1073/pnas.0909075107 Kempton et al.

SUSTAINABILITY SCIENCE Fig. 6. (A) Capacity factors for 11 stations in May 1999 and November 1999; line colors match station colors in Fig 1. The lowest graph, P grid , is the aggregate CF if all stations are connected by transmission. The two gray date ranges are each expanded to 4 d on the maps (B). (B) Sea level pressure (lines, in hPa) and wind speed (color scale at top, in m s1 ) for the events from May 1-4 (Left) and November 4-9 (Right), 1999.

Relating our results to the prior studies reviewed earlier, one (8) found that the rate of decorrelation with distance was related component of the power leveling is the motion of weather systems to the orientationtheir case, the Central Great Plains, shows along the north-south orientation of the Atlantic Transmission quicker decreases in correlation in the east-west direction than Grid. From this perspective, we note that Simonsen and Stevens in north-south, consistently with typical east-west passage of fronts Fig. 7. (A) Capacity factors in June 1998 for 11 stations and for the Grid (Lowermost). The two gray date ranges are expanded to the right. (B) Sea level pressure (lines, in hPa) and wind speed (color scale, m s1 ) for the event of June 13-16 (Left) and June 21-27 (Right).

Kempton et al. PNAS Early Edition 5 of 6

in that region. Conversely, the lack of benefit seen by aggregating developers prospect for the windiest single site, we would advo-stations in the United Kingdom (10) may be due in part to the cate a broader analysisto optimize grid power output by coor-roughly north-south orientation of the island, thus experiencing dinated meteorological and load analysis of an entire region.

their east-west passage of frontal systems nearly simultaneously. This approach to choosing and interconnecting sites has institu-From this regional meteorological perspective, a subsequent tional implications. Today, generation of electricity is primarily a analysis could build on our approach of meteorologically chosen state matter, decided by state public utility commissions, whereas transmission but optimize site choice rather than taking evenly the Independent System Operators (ISOs) manage wholesale placed met stations as we have done. A deliberately optimized power markets and plan transmission. An ISO is the type of orga-array of offshore generator locations should produce even more nization that might plan and operate the electric system we envi-level output, and even fewer times of low power. sion, probably with a mix of ownersprivate firms, existing electric utilities, and/or public power authorities. Because of the unique Discussion characteristics of building and operating offshore, and because In the study region, using our meteorologically designed scale and our proposed Atlantic Transmission Grid would exist primarily orientation, we find that transmission affects output by reducing in federal waters and bridge many jurisdictions on land, it may variance, slowing the rate of change, and, during the study period, make sense to create a unique ISO, here dubbed the Atlantic eliminating hours of zero production. The result is that electric Independent System Operator. Like existing ISOs, the Atlantic power from wind would become easier to manage, higher in ISO would be responsible for managing and regulating the bulk market value, and capable of becoming a higher fraction of power market along the offshore transmission cable, but with electric generation (thus more CO2 displacement).

jurisdiction matched to the synoptic scale of the resource.

Is transmission an economically practical way to level wind? As Whatever the institutions that ultimately manage this resource, an approximate cost comparison, a total of 2,500 MW of offshore we have shown that the nature of wind power generation is dra-wind generation has been approved or requested by states from De-matically altered by scale and interconnectionand we have laware to Massachusetts (all those shown in Fig. 1, plus the 700 MW New York request for proposals). Connecting them by a 3 gigawatt shown the value of a new way of planning transmission corridors, (GW) HVDC submarine cable would require 350 miles of cable. At designing their alignment based on meteorological patterns at the early European offshore wind capital costs of $4,200/kW and sub- synoptic scale.

marine cable capital costs of $4,000,000/mile, the installed costs of Materials and Methods planned offshore wind generation would be approximately To examine our hypotheses, we chose the Eastern Seaboard of the United

$10.5 billion; the connecting transmission would add $1.4 billion States, a span of nearly 2,500 km in northeast-southwest direction. To study (26). They are matched in capacity, each approximately 3 GW, the effects of a large interconnected wind power array, we use anemometer yet the transmission adds less than 15% to the capital cost of gen- data from dispersed stations (using NDBC data) and we model electrical out-eration. This is in line with the market cost of leveling wind via ex- put from the wind speed at each station (SI Text).

isting generation, currently estimated to add about 10% to the cost The colored symbols in Fig. 1 show the locations of NDBC measurements.

of energy (10% cost adder for wind penetrations up to 20%, then a We selected only times for which we had wind speed data from all 11 sta-higher percentage cost added at higher penetration of wind) (27). tions, thus only 59% of the hours during the 5 yr are included in our data-base. Then we extrapolated wind speed from measurement height to turbine Transmission is far more economically effective than utility-scale height and converted from wind speed to power output using previously electric storage (e.g. pumped hydro), whose capital costs are ap-documented methods (23). More information on these methods is in SI Text.

proximately equal to generation. Athorough cost analysis is beyond the scope of this paper, but these approximate comparisons suggest ACKNOWLEDGMENTS. Delaware Sea Grant SG0707 R/CT-3, R.W. Garvine and that transmission costs are commensurate with the value of leveling. W. Kempton (PIs) supported this project. F.M.P. was also supported by CAPES Our findings have implications for the approach taken to wind (BEX 224203-6). The article was improved by suggestions from PNAS development and choice of wind sites. Whereas todays reviewers.

1. Archer CL, Jacobson MZ (2005) Evaluation of global wind power. J Geophys Res 14. Archer CL, Jacobson MZ (2007) Supplying baseload power and reducing transmission 110:D12110 10.1029/2004JD005462. requirements by interconnecting wind farms. J Appl Meteorol Clim 46:1701-1717.

2. Kempton W, Archer CL, Garvine RW, Dhanju A, Jacobson MZ (2007) Large CO2 15. Holttinen H, Hirvonen R (2005) Power system requirements for wind power. Wind reductions via offshore wind power matched to inherent storage in energy end-uses. Powering Power Systems, ed T Ackermann (Wiley, New York), pp 143-167.

Geophys Res Lett 34:L02817 10.1029/2006GL028016. 16. DeCarolis JF, Keith DW (2006) The economics of large-scale wind power in a carbon

3. Roy SB, Pacala SW, Walko RL (2004) Can large wind farms affect local meteorology?. constrained world. Energ Policy 34:395-410.

J Geophys Res 109:D19101 10.1029/2004JD004763. 17. Giebel G On the benefits of distributed generation of wind energy in Europe. PhD

4. Keith DW, et al. (2004) The influence of large-scale wind power on global climate. Proc thesis (University of Oldenburg), p 104.

Natl Acad Sci USA 101(46):16115-16120. 18. Barry RM, Carleton A (2001) Synoptic and Dynamic Climatology (Routledge, New York

5. North American Electric Reliability Council (2008) Generating Unit Statistical Brochure, and London) p 620.

2003-2007, October 2008 (NERC, Princeton, NJ). 19. Milligan MR, Artig R (1999) Choosing wind power plant locations and sizes based on electric reliability measures using multiple year wind speed measurements. National

6. Kempton W, Tomi J (2005) Vehicle to grid implementation: From stabilizing the grid Renewable Energy Laboratory p 9 NREL/CP-500-26724.

to supporting large-scale renewable energy. J Power Sources 144(1):280-294 10.1016/

20. Cassola F, Burlando M, Antonelli M, Ratto CF (2008) Optimization of the regional j.jpowsour.2004.12.022.

spatial distribution of wind power plants to minimize the variability of wind energy

7. Kahn E (1979) The reliability of distributed wind generators. Electr Pow Syst Res 2:1-14.

input into power supply systems. J Appl Meteorol Clim 47:3099-3116.

8. Simonsen TK, Stevens BG (2004) Regional wind energy analysis for the Central United

21. Barthelmie RJ, Courtney MS, Højstrup J, Larsen SE (1996) Meteorological aspects of States. Proc Global Wind Power (American Wind Energy Association, Chicago) p 16.

offshore wind energy: Observations from the Vindeby wind farm. J Wind Eng Ind

9. Czisch G, Ernst B (2001) High wind power penetration by the systematic use of smooth-Aerod 62:191-211.

ing effects within huge catchment areas shown in a European example. Windpower

22. Garvine RW, Kempton W (2008) Assessing the wind "eld over the continental shelf as a 2001 (American Wind Energy Association, Washington, DC). resource for electric power. J Mar Res 66(6):751-773.

10. Oswald J, Raine M, Ashraf-Ball H (2008) Will British weather provide reliable electri- 23. Pimenta F, Kempton W, Garvine R (2008) Combining meteorological stations and city?. Energ Policy 36:3202-3215. satellite data to evaluate the offshore wind power resource of southeastern Brazil.

11. Ernst B, Wan Y, Kirby B (1999) Short-term power fluctuation of wind turbines: Looking Renew Energ 33:2375-2387.

at data from the German 250 MW measurement program from the ancillary services 24. Bahrman MP, Johnson BK (2007) The ABCs of HVDC transmission technologies: An overview viewpoint. Windpower 99 Proceedings; June 20-23, 1999; Burlington, Vermont of high voltage direct current systems and applications. IEEE Power Energy M March/April.

(American Wind Energy Association, Washington, DC). 25. Mesinger F, et al. (2006) North American Regional Reanalysis. Bull Amr Meteorol Soc

12. Milligan MR, Factor T (2000) Optimizing the Geographic distribution of wind plants in 87:343-360 Data at: http://www.emc.ncep.noaa.gov/mmb/rreanl/.

Iowa for maximum economic benefit and reliability. Wind Eng 24(4):271-290. 26. Milborrow D (2009) No consensus on offshore costs. Windpower Monthly Sept:7-9.

13. Sinden G (2007) Characteristics of the UK wind resource: Long-term patterns and 27. Milligan MK, et al. (2009) Wind power myths debunked. IEEE Power Energy M 7 relationship to electricity demand. Energ Policy 35(1):112-127. (6):89-99.

6 of 6 www.pnas.org/cgi/doi/10.1073/pnas.0909075107 Kempton et al.