This atlas estimates wind energy resource for the United States and its territories,(a1). and indicates general areas where a high wind resource may exist. This information is valuable to wind energy developers and potential wind energy users because it allows them to choose a general area of estimated high wind resource for more detailed examination. A siting document, such as that written by Hiester and Pennell (1981), can assist a potential user in going from wind resource assessment to site selection.
The national wind resource assessment was one of the initial goals of the Federal Wind Energy Program. Early research in wind characteristics included the development and application of techniques for estimating the magnitude and distribution of wind resource over a selected area. In 1979 and 1980, the Pacific Northwest Laboratory (PNL) used these resource assessment techniques in preparing twelve regional wind energy atlases covering the United States and its territories (Map A-l and Table A-l). The atlases depicted annual and seasonal average wind resource on a regional and state level. They also included the wind resource's certainty rating and the areal distribution (percentage land area suitable for wind energy development) based on variations in land-surface form. In addition, summary national wind resource maps were produced (Wind Energy Maps 1982) based on a synthesis of the 12 regional assessments (Elliott and Barchet 1981).
A wide variety of data types and analysis techniques were utilized in performing the regional wind energy assessments. Appendix A gives a complete description of the data sources and methodologies used in the regional assessments and their synthesis.
A wind energy data base containing detailed wind statistics for 975 stations in the United States was produced specifically for use in wind energy applications (Barchet 1981). This data base, which was used in producing regional wind energy assessments, was transferred to the National Climatic Data Center (Appendices B and C).
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. However, since the later 1970s, hundreds of new sites have been instrumented specifically for wind energy assessment purposes, and many of these have been located in areas thought to have high wind resource but where data were previously not available or were very limited.
In 1983, the U.S. Department of Energy (DOE) initiated a program administered by PNL to identify and assimilate new site data that could be useful in verifying or updating the wind resource estimates in many areas of the United States. The Pacific Northwest Laboratory contacted numerous federal, state, and private organizations throughout the country regarding existing or planned wind measurement studies to assess the wind energy resource or evaluate potential wind turbine sites. Hundreds of new sites were identified, many with records of sufficient duration to be useful in verifying or updating the previous wind resource estimates. For example, data were available from the DOE measurement program, at thirty-five potential wind turbine sites. The Bureau of Reclamation, Bonneville Power Administration, Western Area Power Administration, Alternative Energy Institute, and California Energy Commission, to name a few, have been involved with instrumenting numerous sites for wind energy assessment or siting purposes. Other organizations, such as the Tennessee Valley Authority (TVA), have performed updated 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).
New site data were identified and obtained for practically every region of the United States, and the majority of these new data were from areas estimated to have high wind resource in the regional atlases. Data were evaluated from approximately 270 new sites for use in verifying or updating the wind resource estimates. Approximately 200 of these new sites were instrumented specifically for wind energy assessment purposes.
The annual and seasonal average wind power maps were revised, based largely on the examination and analysis of these new site data. Certainty ratings credited to the wind resource were revised, and the areal distribution maps were updated to reflect changes in the wind resource estimates. The identification, screening, and evaluation of the new site data and the procedures used in verifying or updating the wind resource, certainty rating, and areal distribution maps are described in Appendix D. Appendix E summarizes data from the 35 DOE measurement sites, also called "candidate sites."
Chapter 2 presents the updated national maps of the annual and seasonal average wind resource, certainty rating, and areal distribution. The annual and seasonal average wind power maps appear in two forms: analyzed versions of the annual and seasonal average wind resource maps and gridded maps. Both are found in Chapter 2. To prepare the gridded maps (Maps 2-6 through 2-25), the analyzed wind resource maps (Maps 2-1 through 2-5) were divided into grid cells of 1/3° longitude by 1/4° latitude over the contiguous United States. The gridded maps were used to assess the certainty of the wind resource estimates and the areal distribution of the wind resources. Different-sized grid cells were used for Alaska, Hawaii, Puerto Rico, and the Virgin Islands.
The gridded maps of the wind resource given in Chapter 2 do not show some of the smaller scale features that are apparent on the analyzed maps. For this reason, the analyzed wind resource maps show greater detail than the gridded maps, especially in mountainous or coastal areas. However, the digitized maps of the wind resource allow the user to associate the wind power classes for specific grid cells with the certainty rating, land-surface form, or any other relevant quantity for those grid cells.
Chapter 3 presents regional summaries of the updated wind resource estimates (Maps 3-1 through 3-72). For each region, major wind resource areas are identified that have been estimated to have suitable wind energy potential for wind turbine applications. For those areas where little or no change was made from the resource estimate in the regional atlases, the descriptive text was extracted and reproduced here with very little revision. Maps of the annual average wind resource are presented individually for each state (or territory) in the region. Some of the larger states (i.e., Alaska, California, and Texas) are subdivided, whereas some of the smaller states are combined on one map. Each map has a latitude-longitude grid to facilitate locating specific places. In addition, each map shows the names of major cities, mountain ranges, geographical features, and prominent wind energy areas for reference purposes.
The wind resource maps estimate the resource in terms of wind power classes (Table 1-1), ranging from class 1 (the lowest) to class 7 (the highest). Each class represents a range of mean wind power density (in units of W/m2) or equivalent mean wind speed at the specified height(s) above ground. Areas designated class 3 or greater are suitable for most wind turbine applications, whereas class 2 areas are marginal. Class 1 areas are generally not suitable, although a few locations (e.g., exposed hilltops not shown on the maps) with adequate wind resource for wind turbine applications may exist in some class I areas.
The wind power estimates apply to areas free of local obstructions to the wind and to terrain features that are well exposed to the wind, such as open plains, tablelands, and hilltops. Within the mountainous areas identified, wind resource estimates apply to exposed ridge crests and mountain summits.
Local terrain features can cause the mean wind energy to vary considerably over short distances, especially in areas of coastal, hilly, and mountainous terrain. Although the wind resource maps identify many areas estimated to have high wind resource, the maps do not depict variability caused by local terrain features.
This wind resource atlas was not intended to deal with variability on a local scale, but to indicate areas where high wind resource is possible. An example of a high wind resource area where considerable local variability occurs is Altamont Pass, California, an area where thousands of wind turbines have been installed. The national wind resource map depicts this area of high wind resource (which appears very small on the national scale map) but does not indicate the local variability which occurs within the area.
Siting handbooks that provide guidelines on siting small and large wind turbines (Wegley et al.1980, Hiester and Pennell 1981, Pennell 1982) address local terrain effects on the wind resource. For finer wind prospecting, consider the siting strategies described in these handbooks.
The wind resource analysis is based on data (where available) collected at heights of 20 to 60 m (65 to 200 ft) above ground at exposed sites. However, in most areas only near-surface data, 3 to 15 m (10 to 50 ft) above ground, were available for the assessment. Vertical extrapolation to 10 and 50 m (33 and 164 ft) is based primarily on the 1/7 power law (Appendix A) using data from exposed sites. Data available from many locations with measurements from more than one level indicate that, in spite of anomalies caused by terrain complexities and nocturnal jets at some locations, the 1/7 power law is generally appropriate (Appendix D). The 1/7 power law conveniently provides wind power densities at 50 m (164 ft) that are twice those at 10 m (33 ft).
The wind power density incorporates in a single number the combined effect of the frequency distribution of wind speeds and the dependence of the wind power on air density and on the cube of the wind speed. In (Table 1-1), the table of wind power classes (which is repeated on the national wind resource maps), the relationship between the mean wind power density and the mean wind speed assumes a Rayleigh distribution(a2) of wind speeds and sea-level air density. The decrease of air density with altitude requires a higher mean wind speed to achieve a given wind power density. To obtain the same wind power density, the mean wind speed must be about 1% higher than shown in the table for every 304 m (1,000 ft) of elevation above sea level.
(Table 1-2) shows why the annual average wind speed alone may not be a reliable indicator of the annual average wind power density. Data from the three locations listed indicate that the locations have identical mean wind speeds at 10 m (33 ft). However, the actual wind power density, which is based on the frequency distribution of the wind speeds, is substantially different for the three locations, such that each location has a different wind power class. The location in New York has a wind speed distribution which is approximated well by a Rayleigh wind speed distribution. The other two locations do not.
In extreme cases, the use of only the mean wind speed and the Rayleigh distribution to estimate the power density provides a much lower estimate than the actual power density. For example, a site near Ellensburg, Washington, has a mean annual wind speed of 5.2 m/s, which is class 3 wind power (160 W/m2) if the Rayleigh distribution is applicable. However, because the distribution of wind speeds at this site is much broader than that of a Rayleigh distribution, the actual wind power is class 6 (320 W/ m2), or twice that estimated by the Rayleigh distribution.
The complexity of the topography and availability of reliable measurements in the vicinity determined the certainty rating credited to the wind resource estimates for exposed locations. These criteria determined the certainty of the wind resource estimate for each grid cell. The maps show the distribution of certainty ratings ranging from 1 for the lowest degree of certainty to 4 for the highest degree of certainty. These maps, depicting the degree of certainty of the wind resource estimates, should be used in combination with the wind resource maps.
Another factor of interest in interpreting wind power resource estimates is their areal distribution, that is, the percentage of land area represented by a specified wind power class. As the ruggedness of the terrain increases, the percentage of land area well exposed to the wind decreases dramatically. Maps in Chapter 2 show the areal distribution exceeding specified wind power classes. These maps indicate various areas of exposure, from mountainous terrain where only a small fraction of the land (<20%) is well exposed to flat terrain where most of the land (>80%) is well exposed.
(a1) U.S. Territories considered are Puerto Rico, three U.S. Virgin Islands (St. Thomas, St. Croix, and St. John), and several U.S. Pacific Islands, or island groups - Midway, Wake and Johnston Islands, Guam, and the Northern Marianas, Marshalls, and Carolines. References to Pacific Islands or Virgin Islands apply only to these U.S. territories.
(a2) The Rayleigh distribution is an analytical expression of a probability density function of wind speed. It seems to fit many observed wind speed distributions reasonably well, although there are exceptions. The advantage of using the Rayleigh distribution is that it is completely specified by one parameter, the long-term average wind speed.
Table of Contents