The twelve regional wind energy resource atlases were based on data collected before 1979. Most of the data used in the assessments were collected at anemometer heights and locations that were not chosen for wind energy assessment purposes. In many areas estimated to have a high wind resource, the certainty rating of this estimate is low because few or no data were available for exposed locations. Since 1978 many locations have been instrumented specifically for wind energy assessment purposes. Many of these new locations have been places thought to have high wind resource but where previous historical data were not available or were very limited.
Since the late 1970s, numerous organizations around the country have been involved in wind measurement studies to assess wind energy potential or evaluate potential wind turbine sites. Many locations selected as potential wind turbine sites were instrumented by the U.S. Department of Energy (DOE). The Bureau of Reclamation (BuRec), Bonneville Power Administration (BPA), Western Area Power Administration (WAPA), Alternative Energy Institute (AEI), and California Energy Commission (CEC), to name a few, have also been involved with instrumenting numerous sites for wind energy assessment or siting purposes. Other organizations, such as the Tennessee Valley Authority (TVA), have performed wind energy assessments incorporating historical data from many sites that were not previously used in the regional atlases (e.g., historical data collected at TVA facilities).
Data records of sufficient length are now available to evaluate the wind resource at hundreds of new sites throughout the country. In this study, data have been evaluated from approximately 270 new sites for use in verifying or updating of the wind resource estimates in the regional atlases. Approximately 200 of the new sites were instrumented specifically for wind energy assessment purposes.
The measured wind resource at the new sites has been compared with the estimated wind resource from the regional atlases, which were based to a large extent on conventional data from airport and military installations. At sites where the measured wind resource was significantly different from the estimated resource, the site's characteristics and wind data were examined in greater detail, where possible. In many cases, data used in the regional atlases were also evaluated in greater detail to determine possible causes for differences between the wind resource indicated by the new site data and that used in the regional atlas. In cases where the evaluation of the new site data indicated that the annual average wind resource estimate(s) for an area(s) should be updated, further analysis was carried out to determine the extent of the area(s) over which these revisions were applicable. This procedure was repeated on a seasonal basis, using seasonal data where available or applicable. In areas where seasonal data were not available or were very limited (for two seasons or less), the change in the seasonal average wind resource was scaled to the change in the annual average wind resource. However, in some areas, new site data indicated that the seasonal trends were substantially different from those presented in the regional atlases. In these areas, further evaluations of the new site data versus the data used in the regional atlases were carried out to determine possible causes for the differences in the seasonal trends. Where the new site data were determined to be more valid and representative than the data or method used in the regional atlases, the seasonal maps were revised as necessary.
In addition to updating the annual and seasonal average wind power maps, maps of the certainty ratings credited to the wind resource estimates and areal distribution (percentage land area) exceeding specified wind power class were also updated. Certainty ratings were upgraded where new site data from exposed sites, especially at heights to 50 m (164 ft) above ground, provided increased confidence in the wind resource estimates. In a few areas, certainty ratings were lowered where new site data of unknown exposure and/or quality indicated a higher or lower wind power class than previously analyzed. In many areas of the Great Plains and Midwest, where a certainty rating of 4 was previously assigned, the certainty rating was lowered if no data were available at heights of 30 to 50 m (164 ft) above ground to verify the wind resource estimate. New site data show that wind power class at 10 m (33 ft) is not always a reliable indicator of the wind power class at 50 m (164 ft), even at apparently well-exposed sites. Thus, estimates for an area that are based solely on data collected at or near 10 m (33 ft) above ground are not credited with the highest certainty rating in the updated assessment. Certainty ratings over inland areas of the Southeastern plains, from eastern North Carolina southward to Florida and westward to eastern Texas, were upgraded to certainty rating 4 everywhere because of factors (e.g., abundant surface data, tower data, small variability in the wind resource, mostly flat to rolling terrain, high roughness) indicating that this region of the country has low wind energy potential with a high degree of confidence for current wind turbine applications.
Areal distribution (percentage land area) maps were not only updated to reflect the revisions in the updated annual average wind power map, but improved to more realistically reflect coastal areas of high wind resource that were omitted in the areal distribution maps in the regional atlases because of the computational scheme used. Areas where these problems occurred have been identified and corrected to provide a more representative analysis in the updated areal distribution maps. In addition, land areas previously omitted in the regional atlases, such as islands in the Great Lakes and islands off the Northeast coast, have been identified and included on the updated areal distribution maps.
In the remainder of this appendix, we describe the sources of the new site data obtained, the procedures used in the screening and evaluation of these data, and the results of the evaluation. The updated wind power analyses, which are based largely on the evaluation and analysis of these new site data, are presented in Chapters 2 and 3 of this atlas, not in this appendix. Site-specific data, such as site location and average wind speed and power, are not presented for the new sites used in this study, except for the DOE candidate wind turbine sites listed in Appendix E. However, references to where the new site data were obtained are included for most of data sources identified in this study.
Comparisons of the estimated versus the measured wind resource at the new sites are described on a region-by-region basis, not on a site-by-site basis except where reference is made to specific DOE candidate wind turbine sites within a region. Some other aspects of the new site data presented here are tables of the distribution of new sites as a function of the measured and estimated wind resources and their difference, comparisons of the actual wind resource versus that computed using a Rayleigh wind speed distribution, and comparisons of the actual wind resource at 50 m (164 ft) versus that based on an extrapolation from 10 m (33 ft).
An effort was made to identify new wind data from various regions of the United States. Organizations contacted in each region were asked about other possible sources of data in their region. This procedure reduced the chance of omitting a major source of new data. However, it was not the intent of this study to identify all sources of new data. Priority was given to identifying summarized wind data, as opposed to raw data, in regions where high wind energy resource was estimated and where new data were collected specifically for wind energy assessment purposes.
Data from a variety of sources were identified and new data were identified in practically every region of the country. A few organizations contacted had proprietary data. Some organizations had just recently installed wind measurement sites or indicated plans to install sites. Several organizations had performed wind energy resource assessments for specific areas incorporating historical wind data used in the regional wind energy atlases and other historical or new data.
The largest amount of data available was from an ongoing project by the California Energy Commission to establish a wind data base of all information acquired within and near the state of California (Waco and Wurst 1983). The data base consists of mean wind speed data acquired by utilities, Federal agencies, and private wind turbine developers. Other major sources of wind data were information acquired from recently completed or ongoing Federal measurement programs (Mah1983; Harrison and Hightower 1983; Bureau of Reclamation 1982, 1983; Baker et al. 1981; Baker, Wade and Persson 1982; Baker and Hewson 1980; Sandusky et al. 1983), and analysis of data available within the Tennessee Valley Authority service region (McGrew, Nielsen and Wiesner 1981). Other data sources were anemometer loan programs (Gipe 1982; Reynolds 1981; Theisen 1983), measurement and data analysis programs sponsored by state energy offices (Sforza, Bailey and Smorto 1980; Otawa, Shoen and Justham 1982; Takle, Brown and Davis 1978), measurement and data analysis programs conducted by universities (Johnson 1982; Myers and Thomann 1982; Lockwood et al. 1981; Martner 1981; Meyers 1979; Huxoll and Wagner 1981; Wagner and Meyers 1981; Ramage, Oshiro and Yokogawa 1979; Wentink 1981; Griscom, Collins and Seavy 1982), and measurement programs conducted by utilities (Kuffel 1982; Colyn and Thero 1982). In total, data from approximately 1100 locations were identified, but after screening only slightly more than 25%, or data from 272 locations, were used in this study.
The format of the obtained data varied markedly. For many locations only mean wind speed data were available, while for others mean wind speed and wind power density were available. The estimates of wind power density reported were usually based on the average of the hourly data or computed from the frequency distribution of the hourly wind speeds. However, the reported values of wind power density from a few sources were computed using an assumed frequency distribution of wind speeds, such as the Rayleigh distribution.
Anemometer heights varied considerably among the sources and frequently among the sites from a given source. At some sites, data were available at two or more levels above ground level. For both the single and multilevel sites, the lowest sensor level was typically at 9 to 15 m (50 ft) above ground level but ranged from 4 to 30 m (98 ft) above ground. Most sources provided summarized data for the heights at which the data were collected; however, a few sources only reported information adjusted to a common height(s) above ground.
The sites for which data were provided were usually assumed to be well exposed to the local wind flow regime. Only a few data sources provided a detailed description of the exposure at each site. Most sites were located in areas estimated to have class 3 or greater wind energy potential in the regional wind energy atlases. One exception was the resource assessment of the Tennessee Valley Authority service area (McGrew, Nielsen and Wiesner 1981), where the majority of sites were located in areas estimated to have only class 1 or 2 wind power.
Once the data were obtained, but before the analysis was begun, data was
screened to eliminate redundancy or data with an insufficient period of record.
In addition, because of the varied formats of the data sets, preliminary
data evaluation and analysis were required. Results of this screening were
used to determine which data would be used for comparison with the results
from the regional atlases. For example, intercomparison of data from two
sites within the same grid cell have resulted in additional data being eliminated
from further comparative studies. Aspects of these tasks are described
Screening of the Data
As a result of the screening process about 75% of the data made available were eliminated for use in the study. Most of the data eliminated fell into one or more of the following categories:
Most data that were eliminated fell into the first two categories. For example, data provided in the wind resource assessment of the BPA service area (Baker, Wade and Persson 1982), the TVA service area (McGrew, Nielsen and Wiesner 1981), California (CEC), New York (Sforza, Bailey and Smorto 1980), and Hawaii (Ramage, Oshiro and Yokogawa 1979) contained a large amount of data previously used in the regional wind energy atlases. Huxoll and Wagner (1981) reanalyzed wind data from the coastal region of Texas but used an extrapolation technique dependent on atmospheric stability. The measurement program being conducted by the Western Area Power Administration (Mah 1983) had only recently begun and only about half of the sites had an adequate amount of data. Less than 10 months of data were available for the sites in New Hampshire (Lockwood et al. 1981), Rhode Island (Griscom, Collins and Seavey 1982), and the anemometer loan program sites in Pennsylvania (Gipe 1982).
Some data were eliminated because the data or site information showed strong evidence of poor exposure. For example, sites with considerably lower wind speeds, in comparison with other data in the vicinity, were eliminated. Surprisingly, data from some of the anemometer loan program sites showed considerably lower wind speeds than existing data from nearby airports and were, thus, eliminated for use in this study.
Analysis and Evaluation of Data
After the initial screening, the data were coded into a standard format. Information coded included: source identification, site number, site name, state, site location by latitude and longitude, site elevation above sea level, height(s) of the anemometer(s) above ground, period of record, mean wind speed, mean wind power density (if available) and the height(s) at which the wind speed and wind power density apply.
Since the data sets existed in various formats, the coded data were first processed so all data were of the same units, and data were adjusted to a common height above ground. Obtaining mean wind speed and available power at the 10- (33 ft) and 50-m (164 ft) levels above the site elevation generally required the use of a suitable extrapolation technique. Mean wind speeds and powers were extrapolated from the anemometer height level to 10 (33 ft) and 50 m (164 ft) above ground by use of the 1/7 power law. For those sites with only mean wind speed data, the available power was estimated by assuming a Rayleigh distribution of wind speeds.
Once the analysis was performed, the wind power classes were determined. These are the same power class ratings used in the regional atlases and this national assessment. Locations of the sites were evaluated to determine if more than one site was within a grid cell (the grid cell dimensions were 1/4° latitude by 1/3° longitude.) If so, the site with the highest power class rating in the grid cell was usually retained unless the data from another site in the cell was considered more representative of mean conditions for an exposed location in the cell (e.g., longer period of record).
After the final screening to eliminate sites within the same grid cell, 270 sites were retained for further evaluation and comparison with the estimates of the wind energy resource in the regional atlases. Table D-l lists the major sources of new data from which at least 10 sites were retained for this study. These sources accounted for about 90% of the new data used in this evaluation. Various other sources accounted for the remaining 10% of the new data used. The site data were dispersed over large geographical areas and primarily located in areas estimated (predicted) to have class 3 or greater wind energy potential, except for the data in the TVA service area and parts of California and New York.
Most of the new sites in the Great Plains from Texas northward to North Dakota were in areas estimated to have class 3 to class 5 wind resource in the regional atlases. New sites in Wyoming and western Montana were also located mostly in areas estimated to have high wind energy potential. In Washington and Oregon, the highest concentration of new sites was in the Columbia River corridor along the Oregon-Washington border, where estimates of the wind resource ranged from class 3 to class 6. In California, new sites were dispersed throughout the state over areas of high and low wind resource. Many new sites were located in or near the Coastal Range wind corridors (e.g., San Gorgonio, Altamont, and Pacheco passes and Carquinez Straits) estimated to have high wind energy potential. Several sites located on mountain summits or ridge crests in Nevada and northeastern California were estimated to have class 5 to 7 wind energy potential. In the East, new sites along the Northeast coast from Long Island to Cape Cod were located in areas estimated to have class 4 and 5 wind resource. Also, several new sites were located along Lake Ontario and Lake Michigan where class 3 and 4 power were predicted. Along the south Texas coast, where a band of class 4 power was estimated, one new site was located on an offshore island and one new site several kilometers inland from the inner coastline.
A comparison of the geographical distribution of the estimated versus measured wind energy resource, adjusted to 10 m (33 ft) at the 270 locations indicated considerably greater spatial variability in the measured resource than in the estimated resource, except over the TVA service area, where uniformly low wind resource was prevalent. Although numerous sites in the Great Plains measured class 4 or greater wind power, quite a few sites in the Great Plains measured only class 1or 2 power. Throughout the West there was considerable intersite variability in the measured wind resource.
Many of these new sites were not installed specifically for wind energy purposes but were installed by or for utilities or other organizations to collect data for other purposes. A few sites were airports or Federal facilities with historical data that were not identified or used in the regional wind energy assessments. Of the 270 sites, 196 sites were determined to have been installed specifically for wind energy assessment purposes in mind. (There was some uncertainty as to the purpose of a few sites.) Approximately two-thirds of the 196 special sites measured class 3 or greater wind resource, whereas roughly one-third measured only class 1 or 2 wind resource. Of the 74 sites that were not installed for wind energy purposes, 80% showed low wind energy potential (class 1 or 2).
Table D-2 gives the number of new sites at which the measured wind resource and at which the estimated wind resource was in each of the seven wind power classes for all new sites and special new sites installed specifically for wind energy assessment purposes. In both cases, the number of sites at which the measured resource is either low (class 1 or 2) or very high (class 6 or 7) is substantially greater than the number of sites predicted. Consequently, the number of sites that had class 3 to 5 power is considerably less than that predicted by the national assessment. In conclusion, there is much more variability in the distribution of the measured resources than in the estimated (predicted) resources.
Considering all 270 new sites, 70% were located in grid cells estimated (predicted) to have class 3 or greater wind power. The measured wind resource was class 3 or greater at 52% of the sites. However, 21% of the sites measured class 5 or greater power, whereas only 13% of the sites were predicted to have this much power.
Considering the 196 special new sites, 86% were located in grid cells estimated to have class 3 or greater power, while 64% measured class 3 or greater power. In areas estimated to have class 5 or greater power, 27% of the sites measured this much power whereas only 16% were predicted to have this power.
Comparisons of Estimated and Measured Resource
The wind resource measured at the new sites was compared with the resource estimates from the regional atlases to determine the difference between the measured and estimated resource at each of the 270 sites. The results of this comparison are not shown here on a site-by-site basis, as it is not the intent of this evaluation to present the site-specific data. Moreover, data from a specific site may not be representative of the general area and could be misleading, without the appropriate information on its location and exposure. Instead, the results of this comparison are described on a region-by-region basis, with some reference to specific sites where examples are used. A summary of the results based on all the new site data in the United States is given at the end of this section.
Over the Great Plains from Texas north to North Dakota and Montana, there were considerably more sites where the estimated resource exceeded the measured resource than vice versa. At some of these sites where the measured resource is less than estimated, we suspect that the site exposure may not be optimum (e.g., local obstructions such as trees and buildings may exist in the vicinity). It appears that some of the sites had lower wind resource than estimated because they are located in areas of relatively lower elevation than nearby terrain. A few of the sites that were located near airports had considerably lower wind speeds than the nearby airport during the same period, indicating that the site's anemometer was not well exposed.
However, a large fraction of the new sites throughout the Great Plains had the same wind power class as estimated or only slightly greater or less than that estimated. This indicates that, for the most part, the estimates in the regional atlases are fairly representative of the typical exposed locations in the Great Plains.
Ten sites in the Great Plains had considerably greater wind resource than estimated. Most of these sites were located on elevated terrain features, which were higher than much of the surrounding terrain. The regional atlases depict the prominent ridge crests and mountain ranges in the United States but do not depict less prominent terrain features such as the relatively low ridges and hills scattered throughout portions of the Great Plains. However, many of these elevated terrain features in the Great Plains can be expected to have greater wind energy potential than that estimated for the open plains and rolling country, as was indicated in the assessment of the North Central region (Freeman et al. 1981).
The review of all the new wind data for the Great Plains from Texas to the Dakotas indicates several areas where the representative, new site data indicate higher or lower wind resource than was estimated in the regional atlases. One such area was the class 5 area in the southern Great Plains over the Texas-Oklahoma Panhandle. Eight new sites were located in this region, including four sites that also had data at approximately 50 m (164 ft) above ground. None of these new sites indicated the class 5 wind power at either the 10-m (33 ft) or 50-m (164 ft) level that was estimated in the regional atlases, but instead measured class 3 and 4 power. In the regional atlases, the data used in this area were older airport data from the 1930s to the early 1950s as no recent data were summarized or digitized. The authors now believe that these older data are biased on the high side. For example, at Clayton, New Mexico, the airport data from 1948 to 1951 show wind power class 5 (280 W/m2 at 10 m or 33 ft), whereas the nearby new site data from 1977 to 1982 at Clayton give wind power class 3 (170 W/m2 at 10 m or 33 ft). Both site locations appear almost equally well exposed to the wind. An interesting note is that there is only a 1 m/s difference in the mean annual wind speeds at 10 m (33 ft) between these two locations, but this results in a significant difference in the wind power density (over 100 W/m2 difference at 10 m or 33 ft).
New site data indicate that the estimates may be on the low side in parts of North Dakota. Two years of new data were collected at 9 m (30 ft) and 45 m (148 ft) near Finley in eastern North Dakota. The low-level data indicated class 4 power, which agrees well with the estimates in the regional atlases. However, at the upper level the wind power was class 6 (740 W/m2 at 45 m or 148 ft) at this site. A more detailed analysis of this site's data reveals very strong nocturnal shear, which results in considerably greater power at 50 m (164 ft) than at 10 m (33 ft). If this site's wind regime is characteristic of that over the larger areas of eastern North Dakota and western Minnesota, then the wind power estimates in the regional atlas may be one to two power classes too low (for the 50 m [164 ft level]). Additional data are needed to more reliably estimate the extent of the wind resource over these areas.
In western North Dakota, new site data at five different exposed locations indicate class 4 to 5 wind power potential at 10m (33 ft), in comparison to the class 3 to 4 estimates in the regional atlases. A site near Minot indicates class 5 potential at 50 m (164 ft) for exposed areas in western North Dakota, in comparison to the class 4 power estimated for 50 m (164 ft) in the regional atlases.
In southern Wyoming, 16 new sites indicate considerable variability in the wind resource, ranging from class 1 to class 7. A few of these new sites with low wind resource are suspected of having poor site exposure. Data from exposed sites in southeastern and southwestern Wyoming indicate class 6 wind power at 10 m (33 ft) and 50 m (164 ft), in comparison to the class 4 and 5 estimates in the regional atlases.
In the TVA service area in the Southeast, additional site data confirm, to a higher certainty, that the resource is low throughout this region, except for exposed mountain summits in the Appalachians. In the Northeast, many sites had lower wind resource than estimated; however, many of these were sites that were not installed for wind energy purposes. Thus, the exposure of these sites is questionable. Exposed sites along the Atlantic coast from Long Island to Cape Cod and along the coasts of Lake Ontario and eastern Lake Erie mostly had comparable or slightly less wind resource than indicated by the estimates in the regional atlases.
In mountainous regions (as in much of the western U.S.), comparisons of the measured and estimated wind power classes at specific sites can be misleading if the types of terrain features represented by the grid cell estimate and site location are not the same. For example, the grid cell estimate may be representative of a ridge crest or mountain summit, whereas the site location may be in a valley. Likewise, the opposite situation may occur where the site is on an exposed ridge crest, but the estimate is representative of a plain or valley.
However, in most cases the grid cell estimate and new site data are for locations of the same terrain type (i.e., ridge crest, broad valley, open plain, tableland, etc.). In complex terrain, many of the large differences between the estimated and measured resource may be attributed to local variations in the resource over the same terrain feature(s) within the same grid cell. For example, the wind resource at 10 m (33 ft) along a ridge can vary by several wind power classes from one part of the ridge to another part. This has been documented in numerous studies of the resource in complex terrain areas.
Table D-3 gives the number of sites in the United States for which the difference in wind power class (measured minus estimated) was a given amount. At 65% of all new sites, the measured wind power was within one power class of the estimated power, and at 30% of these there was no difference in the power class. The measured wind power at 35% of all new sites differed by 2 or more classes from the estimated power; approximately half of these were greater and half less than the estimated power class. For special new sites (those sites installed specifically for wind energy purposes), 26% of the sites had the same class as the estimated power and 57% of the sites were within ± 1 power class. Thus, 43% of the sites differed by 2 or more power classes from the estimated power class.
Table D-4 shows the distribution of the number of new sites by measured wind power class and the difference between the measured minus the estimated wind power class. For new sites that have low wind resource (power class 1 or 2), the measured resource is less than the estimated resource at 67% of the sites and greater than the estimated resource at only 5% of the sites. However, for new sites with class 4 or greater wind resource, over 60% had higher wind resource than estimated whereas only 9% had less resource than estimated. Approximately 65% of the sites with class 5 or greater resource had considerably more wind resource (two or more classes greater) than estimated. Thus, the information in Table D-4 indicates that a significant percentage of the sites with high wind resource were in areas estimated to have much lower wind resource, and vice versa.
At 109 of the 270 new sites, only mean wind speed data were provided. At these sites, the wind power was computed assuming a Rayleigh distribution of wind speeds.
At the other 161 sites, the wind power was based on the actual distribution of wind speeds. For these sites, the wind power based on a Rayleigh speed distribution was also computed and compared with the actual wind power. The results of this comparison are shown in Table D-5. These results indicate that use of the Rayleigh distribution would underestimate the power class at 37% of the sites and overestimate the power class at only 4% of the sites.
If we consider only those sites where moderate-to-high wind power (class 3 or greater) was measured, then the results are quite different. The wind power was class 3 or greater at 75 sites. Use of the Rayleigh distribution would underestimate the wind power class at 52% of these sites, and at 15% of the sites the power would be underestimated by two or more power classes. The power would be overestimated by one class at only 7% of the sites. At none of the sites would the power be overestimated by two or more classes.
Thus, the data available for this study indicate that in areas of high wind energy potential, use of the Rayleigh distribution frequently underestimates the power but rarely overestimates the power. These results apply to 10 m (33 ft) and represent an average over a wide geographical distribution of sites throughout the contiguous United States. They may not apply to certain specific regions or be entirely representative of any given region.
For the majority of the 270 new sites, only one level of wind data was available and it was usually nearer the 10-m (33 ft) level than the 50-m (164 ft) level. The wind power at 50 m (164 ft) at these sites was estimated using the 1/7 power law equation.
However, at 63 new sites wind data were available for two or more heights above ground, usually near the 10-m (33 ft) and 50-m (164 ft) levels. At these sites, a more accurate estimate of wind power at the 50-m height could be obtained than at sites with only one level of data near 10 m (33 ft). A comparison was made between the actual wind power at 50 m (164 ft), based on data collected at or near the 50-m (164 ft) level, and the estimated wind power at 50 m (164 ft) based on data collected at or near 10 m (33 ft) and extrapolated to 50-m (164 ft) using the 1/7 power law. The results of this comparison are shown in Table D-6 for sites with class 1 or 2 wind power and for sites with class 3 or greater wind power.
At 38 sites where class 3 or greater wind power was measured at or near the 50-m level, only 16 sites (42%) would have the same wind power class at 50 m (164 ft) if 10 m (33 ft) data and the 1/7 power law were used to estimate the 50-m (164 ft) power. The actual wind power class at 50 m (164 ft) was greater than the estimated power class at 37% of the sites, and at half of these the actual exceeded the estimated by two or more power classes. Most of the sites with the largest difference between the actual and estimated power classes at 50 m (164 ft) are located in areas where trees are prevalent in the surrounding environment. For example, at many of the sites in the eastern United States (e.g., Northeast and Great Lakes regions), the actual wind power at 50 m (164 ft) is considerably greater than the estimated power.
Over most of the Great Plains, there was little difference in the actual and estimated power at 50 m (164 ft), so the 1/7 power law appears appropriate to most exposed locations throughout the southern and central Great Plains. An exception to this is eastern North Dakota, where the actual wind power at 50 m (164 ft) was two power classes greater than the estimated power. Very strong nocturnal shear primarily accounts for this large difference between the actual and estimated power at 50 m (164 ft) in this area.
The actual power at 50 m (164 ft) was less than the estimated power at 21% of the 38 sites with moderate-to-high wind power at 50 m (164 ft). Most of these sites are located on ridge crests, hilltops, and other elevated terrain features where a local acceleration of the wind is caused by the terrain feature. For example, the actual wind power at 50 m (164 ft) was considerably less than the estimated power extrapolated up from 10 m (33 ft) at Wells, Nevada (a ridge crest site), San Augustin Pass, New Mexico (a mountain pass), and Point Conception, California (a coastal head).
Baker, R. W., and E. W. Hewson. 1980. Network Wind Power Over the Pacific Northwest: Progress Report, October 1979 - September 1980. BPA 80-5, Bonneville Power Administration, Portland, Oregon.
Baker, R. W., J. E. Wade, P.O.G. Persson, and B. Armstrong. 1981. Regional Wind Energy Assessment Program Progress Report, October 1980-September 1981. BPA 81-6, Bonneville Power Administration, Portland, Oregon.
Baker, R. W., J. E. Wade, and P.O.G. Persson. 1982. Regional Wind Energy Assessment Program Progress Report, October 1981 - September 1982. BPA 82-8, Bonneville Power Administration, Portland, Oregon.
Bureau of Reclamation. 1982. Central California - Nevada Wind Energy Study, U.S. Department of the Interior, Bureau of Reclamation, Sacramento, California.
Bureau of Reclamation. 1983. Northern Great Plains Wind Energy Study- Special Report, U.S. Department of the Interior, Bureau of Reclamation, Billings, Montana.
Colyn, C., and M. Thero. 1982. Wind Resource Assessment Project- Third Quarter, 1982. Tri-State Generation and Transmission Association, Inc., P.O. Box 33695, Denver, Colorado.
Freeman, D. L., D. L. Hadley, D. L. Elliott, W. R. Barchet, and R. L. George. 1981. Wind Energy Resource Atlas: Volume 2 - The North Central Region. PNL-3 195 WERA- 2, Pacific Northwest Laboratory, Richland, Washington.
Gipe, P. 1982. Adopt an Anemometer- A Unique Anemometer Loan Program. The Center for Alternative Resources, P.O. Box 539, Harrisburg, Pennsylvania.
Griscom, C. A., C. Collins, and G. Seavey. 1982. Coastal Rhode Island Wind Resource Analysis Project 19811982 - Volume 1. Division of Marine Resources, University of Rhode Island, Narragansett, Rhode Island.
Harrison, W. F., and S. J. Hightower. 1983. Wind Site Prospecting Studies, U.S. Department of the Interior, Bureau of Reclamation, Denver, Colorado.
Huxell, V. F., and N. K. Wagner. 1981. Stability Adjusted Wind Power in the Texas Coastal Zone. Report 56, Atmospheric Science Group, College of Engineering, The University of Texas, Austin, Texas.
Johnson, G. L. 1982. Kansas Wind Resource Assessment. July, 1980 - June, 1982, Report 151, Engineering Experiment Station, Kansas State University, Manhattan, Kansas.
Kuffel, L. 1982. Wind Energy Research and Demonstration Program, Wind Characteristics - Tall Tower Study, Wisconsin Power and Light Company, Madison, Wisconsin.
Lockwood, J., G. Kraft, G. Pregnent, and L. Smukler. 1981. Study of a Wind Energy Conversion System in New Hampshire, University of New Hampshire, Durham, New Hampshire.
Mah, D. J. 1983. Wind Measurement Activity D-83-A24-WIND-0001, Western Area Power Administration, P.O. Box 3402, Golden, Colorado.
Martner, B. E. 1981. Wind Characteristics in Southern Wyoming, Part 1: Surface Climatology, Report AS 127. Department of Atmospheric Science, University of Wyoming, Laramie, Wyoming.
McGrew, D., R. Nielsen, and W. Wiesner. 1981. Wind Resource Analysis for the Tennessee Valley Authority Region - Volume 1. Boeing Engineering and Construction Company, Seattle, Washington.
Meyers, C. E. 1979. Wind Speecl and Wincl Pov~er Potential in a Coastal Environment. Report No. 52, Atmospheric Science Group, U niversity of Texas, Austin, Texas.
Myers, G. A., and G. C. Thomann. 1982. "Wind Resource Study of the Great Plains." Preprints from Wind/Solar Technology Conference, Kansas City, Missouri.
Otawa, T., D. A. Shoen, and S. A. Justham. 1982. Indiana Wind Energy- A Guide to Harnessing Hoosier Wind Power, Ball State University, Muncie, Indiana.
Ramage, C. S., N. E. Oshiro, and S. T. Yokogawa. 1979. Hawaii Wind Power Surver: Fixed Station Data, Report UHMET 79-15, Department of Meteorology, University of Hawaii, Honolulu, Hawaii.
Reynolds, R. D. 1981. State of New Mexico Wind Site Survey Loan Program, Report 2-68-2216, Physical Science Laboratory, New Mexico State University, Las Cruces, New Mexico.
Sandusky, W. F., J. W. Buck, D. S. Renne, D. L. Hadley, O. B. Abbey, S. L. Bradymire, and J. L. Gregory. 1983. Candidate Wind Turbine Generator Site Cumulative Meteorological Data Summary and Data for January 1982 Through September 1982. PNL-4663, Pacific Northwest Laboratory, Richland, Washington.
Sforza, P. M., B. H. Bailey, and M. J. Smorto. 1980. High Yield Wind Energy Resources in New York State, NYSERDA 80-11, New York State Energy Research and Development Authority, Albany, New York.
Takle, E. S., J. M. Brown, and W. M. Davis. 1978. "Characteristics of Wind and Wind Energy in lowa." lowa State Journal of Research 52(3):313-339.
Theisen, P. M. 1983. Anemometer Loan Program Year End Report - July 1983. Wisconsin Division of State Energy, Madison, Wisconsin.
Waco, D. E., and M. P. Wurst. 1983. "California's Wind Measurement Program." Paper presented at the Sixth Biennial Wind Energy Workshop, Minneapolis, Minnesota, June 1-3, 1983.
Wagner, N. K., and C. E. Meyers. 1981. Coastal Wind Assessment: Interviews and Observations. Atmospheric Science Group, University of Texas, Austin, Texas.
Wentink, T. 1981. Winds at an Interior Alaska Summit. Geophysical Institute, University of Alaska, Fairbanks, Alaska.
Table of Contents