close

Вход

Забыли?

вход по аккаунту

?

990

код для вставкиСкачать
INTERNATIONAL JOURNAL OF CLIMATOLOGY
Int. J. Climatol. 19: 471–488 (1999)
CLIMATOLOGICAL MEAN AND INTERANNUAL VARIANCE OF
UNITED STATES SURFACE WIND SPEED, DIRECTION AND
VELOCITY1
KATHERINE KLINK*
Department of Geography, 414 Social Sciences Building, Uni6ersity of Minnesota, Minneapolis, MN 55455, USA
Recei6ed 24 February 1998
Re6ised 14 August 1998
Accepted 3 September 1998
ABSTRACT
Means and variances of monthly mean wind speed, direction and velocity (the mean resultant vector) are derived for
the period 1961–1990 at 216 stations in the coterminous United States. Direction and velocity means and variances
are calculated using a complex-arithmetic extension of the equations for scalar mean and variance. Variance is derived
from the 30-year time series of monthly means. While analyses of monthly mean wind fields are common,
accompanying analyses of speed, direction, and velocity variance do not generally accompany them.
Mean monthly wind direction and velocity fields show a typical seasonal progression from westerly and
northwesterly winds in winter, to southerly winds in summer. Scalar and vector wind speeds are highest in winter and
spring, and lowest in the summer. Seasonal variation in the mean fields is related to seasonal changes in mean sea
level pressure, particularly east of the Rocky Mountains. In the western United States, mean winds often reflect
channeling by local topography. The interannual variance of mean monthly wind speed, direction and velocity are
related to seasonal variability in synoptic-scale features, such as the frequency of cyclones and anticyclones. Low
variance occurs at a number of stations in the west, where topography restricts the range of wind variability. High
velocity variance appears when both speed and direction variability are high, but it can occur also when speed
variance is high and direction variance is low (or 6ice 6ersa). Low velocity variance can result from low speed and
direction variance, or from low mean wind velocities. The mean and variance characteristics of surface winds provide
additional information on the surface climatology of the coterminous United States, and serve as a useful adjunct to
other extant land-surface climatologies. Copyright © 1999 Royal Meteorological Society.
KEY WORDS: wind
speed; wind direction; wind velocity; directional and vector statistics; USA
1. INTRODUCTION
The wind field has been studied in several manners, and for many purposes. Wind speed is used to
evaluate wind power potential (Palutikof et al., 1987), wind direction is a factor in bird and insect
migration (Burt and Pedgley, 1997), and wind velocity (the wind vector) is important for determining
particulate dispersion (Cabezudo et al., 1997). Wind data are also used for applications such as
precipitation gauge correction (Legates and Willmott, 1990) and model validation (Roads et al., 1995).
This research evaluates the monthly mean and interannual variability of surface wind speed, direction
and velocity—the mean resultant wind — at 216 stations in the coterminous United States for the period
1961–1990. (Although sometimes used interchangeably, ‘speed’ and ‘velocity’ here are distinct: speed
refers to a scalar variable and velocity is strictly a vector.) Directional and vector means and variances are
computed using a complex-arithmetic extension of the equations for scalar mean and variance. The
* Correspondence to: Department of Geography, 414 Social Sciences Bldg., University of Minnesota, Minneapolis, MN 55455,
USA. Tel.: +1 612 6253452; fax: +1 612 6241044; e-mail: klink@atlas.socsci.umn.edu
1
Monthly data and station histories are available from the author.
Contract/grant sponsor: Minnesota Supercomputer Institute
CCC 0899–8418/99/050471 – 18$17.50
Copyright © 1999 Royal Meteorological Society
472
K. KLINK
1961–1990 monthly mean sea level pressure field is included for comparison to surface wind speed,
direction and velocity.
This work is not the first climatological analysis of United States surface winds, but it is unique in
providing mean and variance fields together. Several atlases include maps of the climatological means of
wind speed, direction and/or velocity (Visher, 1954; U.S. Department of Commerce, 1968; USGS, 1970;
Elliott et al., 1986), although each source uses different station densities and climatological averaging
periods. Atlases sometimes show only the seasonal average (USGS, 1970) or January and July wind fields
(Visher, 1954). Only rarely does a single atlas include the speed, direction and velocity fields together (the
Department of Commerce atlas cited above is an exception, although the velocity maps are only for
January, April, July and October). It also is unusual to find accompanying maps of the variance of wind
speed, direction or velocity. Wind roses are one means to depict wind direction variability (and in
modified form, wind velocity variability), and the Department of Commerce atlas includes monthly wind
roses for wind directions. However, none the above sources include maps of wind speed or wind velocity
variability.
2. DESCRIPTIVE STATISTICS FOR WINDS
Wind analyses use scalar statistical methods for wind speeds or when wind directions and velocities (wind
vectors) have been separated into their easterly and northerly components. When whole directions or
vectors are included in the analysis, slightly modified statistical methods are required (Legler, 1983;
Oehlert, 1983; Turner, 1986; Fisher, 1987; Breckling, 1989; Klink and Willmott, 1989; Hanson et al., 1992;
Jönsson and Fortuniak, 1995; Feliks et al., 1996; Weber, 1997). Several methods have been used to
compute variances (or standard deviations) for directional data (Mardia, 1972; Gaile and Burt, 1980;
Batschelet, 1981; Fisher, 1993). In this paper, the directional and vector mean and variance are based on
the equivalent scalar equations so that wind speed, direction and velocity statistics are computed using the
same procedures (Klink, 1998).
The mean and variance of a set of wind speed observations are computed using the equations for scalar
mean and variance. For wind directions and velocities, the means and variances are evaluated by
generalizing these equations to accommodate complex numbers:
z̄ =
1 n
% z = x̄ +iȳ
n j=1 j
(1)
and
s 2z =
1 n
% (z −z̄) (zj −z̄) = s 2x +s 2y
n j=1 j
(2)
where zj = xj + iyj is the complex-arithmetic representation of the wind direction or wind velocity for
month j, xj is the east – west component of the wind (direction or velocity) for month j, yj is the
north–south component, i = − 1, and * represents the complex conjugate. The mean direction and
mean velocity have the units of the original observations. Directional variance is a dimensionless number
that ranges between zero (all directions are the same) and one (directions lack a single mode of
concentration; see Figure 1). The velocity variance has units of wind speed squared, a lower limit of zero
and no fixed upper limit.
3. WIND DATA SOURCE
Monthly mean wind speeds, directions and velocities were derived from hourly and three-hourly surface
wind speed and direction observations from 1961 to 1990 (NREL, 1993). Most of the observations were
recorded hourly, but winds were recorded only every 3 h from the mid-1960s to the mid-to-late 1970s
Copyright © 1999 Royal Meteorological Society
Int. J. Climatol. 19: 471 – 488 (1999)
VARIANCE OF UNITED STATES SURFACE WIND SPEED, DIRECTION AND VELOCITY
473
(depending on the station). Relatively few stations had significant gaps in the available record (Table I);
the number of missing hourly or three-hourly observations at a given station for a given month was
generally no larger than 20%, and usually it was much smaller. Wind speed statistics are based only on
the hourly and three-hourly speed observations; direction statistics are computed from wind directions,
disregarding the speed; and velocity statistics are derived from the hourly and three-hourly speed and
direction observations taken together. Climatological means and variances are computed using the 30-year
series of mean monthly values (speed, direction, velocity) for each station.
Station histories (NOAA (1993), supplemented with data from Changery (1978)) showed that over the
period of record the majority of stations in the data set had undergone instrument moves and changes in
measurement heights. Anemometer heights ranged from about 6.1 to 21.3 m (20–70 ft), and many early
anemometers were on the roofs of buildings. During the 1960s, most stations moved their instruments
from rooftop to ground level, and positioned their anemometers at (or near) 6.1 m (20 ft) above the
ground. In the early to mid-1980s, about 25% of the stations repositioned their anemometers at 10 m (33
ft), the World Meteorological Organization standard wind measurement height. The most common height
during the 1961– 1990 period was 6.1 m, and the 1/7 power law (Peterson and Hennessey, 1978) was used
to standardize the hourly and three-hourly wind speed measurements to this height. Speeds were left
unchanged for periods in which no instrument history was available, which involved about 13% of the
stations (Table II). Aside from the changes in anemometer height, instrument moves were assumed not to
affect wind speed readings.
Wind directions (recorded in 10° increments) were assumed to be unaffected by instrument moves or
changes in measurement height. Simultaneous measurements of wind direction at different heights were
not available. Although Ekman theory describes in principle how wind directions will change when
nearing the ground surface, observations rarely match the predictions of the theory (Holton, 1979). No
attempt was made to modify wind directions for changes in instrument height, because directional
correlation statistics (Hanson et al., 1992) computed at several randomly selected stations using directions
recorded before and after instrument height changes showed no consistent angular deviation.
Thirty-two of the 216 stations in the data set had periodic gaps in the wind record and 29 stations had
incomplete station histories (some stations had both). Although data gaps and uncertain height corrections may introduce a bias into the monthly statistics, these stations were retained for the analysis because
deleting them would significantly reduce station density, particularly in the western United States, and
because spatial patterns in the speed, direction and velocity fields showed few obvious anomalies.
Figure 1. Example directional distributions. The data in (a) have a variance of zero, while the data in (b) and (c) both have variances
equal to one (the maximum possible value)
Copyright © 1999 Royal Meteorological Society
Int. J. Climatol. 19: 471 – 488 (1999)
474
K. KLINK
Table I. Wind stations with significant data gaps. Missing data codes are: (1) missing many or all
overnight observations for speed and direction; (2) missing many or all overnight observations for
direction only; (3) missing many or all 24-h observations for speed and direction
WBANa
number
City and state
Dates (month/day/year) and missing data code
03103
Flagstaff, AZ
12912
Victoria, TX
13729
13733
Elkins, WV
Lynchburg, VA
14607
14891
Caribou, ME
Mansfield, OH
14926
14941
23048
St. Cloud, MN
Norfolk, NE
Tucumcari, NM
23061
23063
Alamosa, CO
Eagle, CO
23129
23153
23155
23161
23273
24013
24036
Long Beach, CA
Tonopah, NV
Bakersfield, CA
Daggett, CA
Santa Maria, CA
Minot, ND
Lewistown, MT
24037
Miles City, MT
24128
24137
Winnemucca, NV
Cut Bank, MT
24146
24221
93037
93058
93842
93987
94185
Kalispell, MT
Eugene, OR
Colorado Springs, CO
Pueblo, CO
Columbus, GA
Lufkin, TX
Burns, OR
94702
94746
94814
Bridgeport, CT
Worcester, MA
Houghton, MI
94849
Alpena, MI
01/01/61–10/31/61
01/01/64–12/31/64
01/01/61–06/01/61
07/01/61–06/30/64
07/28/68–12/31/90
01/01/61–06/30/76
09/01/89–12/31/90
08/02/63–12/31/90
01/01/61–09/30/66
08/03/81–12/31/90
01/01/61–12/31/90
01/01/61–06/30/76
10/04/82–12/31/88
01/01/89–12/31/90
01/01/61–12/31/90
06/22/85–12/31/88
01/01/89–12/31/90
02/16/86–12/31/90
02/06/82–06/13/88
10/01/83–12/31/90
09/20/81–12/14/87
06/15/64–12/31/90
01/01/89–12/31/90
03/28/86–07/31/90
08/01/90–12/31/90
04/23/89–12/31/89
01/01/90–12/31/90
01/25/88–12/31/90
11/01/81–07/31/83
04/29/84–12/31/88
01/01/89–12/31/90
01/01/61–06/27/64
01/01/61–06/14/64
01/01/64–12/31/64
03/23/78–12/31/90
01/01/61–12/31/64
01/23/86–01/31/90
01/10/73–12/31/88
01/01/89–12/31/90
06/01/81–12/31/90
07/04/84–12/31/90
04/20/64–09/30/64
07/19/79–02/10/80
01/11/81–12/31/90
01/01/61–05/06/64
10/15/64–11/30/65
a
(1)
(2)
(1)
(1)
(1)
(1)
(1)
(1)
(1)
(1)
(1)
(1)
(1)
(3)
(1)
(1)
(3)
(1)
(1)
(1)
(1)
(1)
(3)
(1)
(3)
(1)
(3)
(1)
(1)
(1)
(3)
(1)
(1)
(2)
(1)
(1)
(1)
(1)
(3)
(1)
(2)
(1)
(1)
(1)
(1)
(1)
WBAN: Weather Bureau Army Navy.
4. MEAN MONTHLY SURFACE WINDS IN THE COTERMINOUS UNITED STATES
4.1. Mean wind 6elocity (mean resultant wind)
East of the Rockies, the annual cycle of wind velocity follows a smooth and fairly predictable cycle
(Figure 2). Westerly and northwesterly winds are dominant in the winter, and are particularly strong in
Copyright © 1999 Royal Meteorological Society
Int. J. Climatol. 19: 471 – 488 (1999)
VARIANCE OF UNITED STATES SURFACE WIND SPEED, DIRECTION AND VELOCITY
475
the northern states. Spring velocities reflect the transitional nature of this season, with many stations
having low or near-zero mean resultant winds (mean velocities). A dominant southerly flow in the central
United States and southern Florida is maintained throughout the summer, but wind velocity is quite
variable in the southeast, particularly in July and August. The autumn transition is essentially the reverse
of the spring pattern, with southerly flow giving way to westerly and northwesterly winds.
Surface wind patterns in the western United States are strongly influenced by the variable topography.
Along the Front Range of the Rockies, wind velocities are very consistent in both magnitude and
direction from October to April (Figure 2). Topographic channeling is a major contributor to the strong
and persistent winds in this region (Martner and Marwitz, 1982). Topographic effects are clearly
dominant at Pocatello, Idaho; Salt Lake City, Utah and Ely, Nevada—each shows little month-to-month
change in the mean resultant wind. In contrast to the pattern for other parts of the country, winds along
the west coast of the United States are strongest from late spring to early autumn, and are lowest during
the winter. Hourly wind records show that day-to-day variability in wind direction is higher, and calm
winds are experienced more often, in winter than in summer. The result is that, when averaged over a
month, winter mean velocities are lower than the summer values.
Table II. Wind stations without complete instrument histories. For gaps early in the period (first group in the Table),
wind speeds were left unchanged prior to the beginning of instrument records. For gaps later in the wind record
(second group), wind measurement heights were assumed to be the same as the latest known height and speeds were
corrected to 6.1 m (20 ft) if necessary. For the three stations without available instrument histories, no height
correction was applied to the data
WBANa number
City and state
Earliest available instrument
history (month/day/year) and
anemometer height
03860
03870
93738
94240
94728
94814
04751
14751
14850
14940
14991
23048
23153
23161
23184
24013
24025
24027
24036
24037
24137
24230
24283
24284
93987
94725
23063
94018
94185
Huntington, WV
Greenville, SC
Sterling, VA
Quillayute, WA
New York, NY
Houghton, MI
Bradford, PA
Harrisburg, PA
Traverse City, MI
Mason City, IA
Eau Claire, WI
Tucumcari, NM
Tonopah, NV
Daggett, CA
Prescott, AZ
Minot, ND
Pierre, SD
Rock Springs, WY
Lewistown, MT
Miles City, MT
Cut Bank, MT
Redmond, OR
Arcata, CA
North Bend, OR
Lufkin, TX
Massena, NY
Eagle, CO
Boulder, CO
Burns, OR
12/01/61; 6.1 m
10/15/62; 7.0 m
11/17/62; 6.1 m
08/01/66; 6.7 m
05/— /62; 20.7 m
04/20/64; 12.8 m
a
Latest available instrument
history (month/day/year) and
anemometer height
07/23/68;
06/01/61;
08/28/62;
03/04/51;
07/17/61;
09/18/59;
07/12/59;
04/18/61;
04/06/65;
06/29/62;
06/07/62;
07/27/60;
10/13/64;
11/07/36;
10/04/59;
09/23/59;
07/30/59;
04/30/59;
04/23/65;
12/09/66;
6.4 m
6.7 m
6.1 m
6.1 m
8.5 m
6.7 m
7.9 m
6.1 m
6.1 m
6.1 m
6.1 m
6.1 m
6.1 m
12.2 m
6.1 m
6.1 m
6.1 m
6.1 m
6.1 m
6.1 m
No information
No information
No information
WBAN: Weather Bureau Army Navy.
Copyright © 1999 Royal Meteorological Society
Int. J. Climatol. 19: 471 – 488 (1999)
476
Copyright © 1999 Royal Meteorological Society
K. KLINK
Int. J. Climatol. 19: 471 – 488 (1999)
Figure 2. Mean monthly wind velocity, 1961 – 1990. The arrows are centered on station locations and fly with the wind. A 2 m s − 1 arrow is shown for scale (December and
November)
VARIANCE OF UNITED STATES SURFACE WIND SPEED, DIRECTION AND VELOCITY
477
4.2. Mean wind speed
For much of the country, monthly mean wind speeds are highest in winter and spring when the
equator-to-pole temperature and pressure gradients are strongest. Many stations report the highest
speeds in March and April (Figure 3). The slowest monthly wind speeds generally occur in July,
August and September, a time of diminished latitudinal temperature and pressure gradients. Wind
speeds are high throughout the year at most stations in the central United States and in the northeast.
In contrast, there is a distinct area in the southeastern United States that has much lower wind
speeds, with stations in this area recording values among the lowest in the country.
In the western United States, stations with high monthly wind speeds are interspersed with stations
having very low values (Figure 3). A few stations have low mean speeds throughout the year (e.g.
Elko, Nevada; Missoula, Montana; Medford, Oregon) while others maintain comparatively high mean
monthly speeds (e.g. stations along the Front Range of the Rockies). In contrast to the central and
northeastern United States, several stations along the west coast (e.g. San Francisco and Santa Maria,
California; North Bend, Oregon) experience their highest wind speeds in late spring and in the
summer, rather than in the winter and early spring.
4.3. Mean wind direction
Monthly maps show that in general wind directions are more spatially coherent east of the Rockies
than in the mountainous west (Figure 4). Winter mean directions in the central and northeastern
United States are westerly and northwesterly, become more variable during the spring, and take on a
dominant southerly component by June. Southerly dominance begins to diminish in early autumn as
directions resume the winter pattern. Winds in the southeastern United States do not follow this
seasonal trend. Winter winds are more spatially variable in the southeast than in the central and
northeastern states, although like those regions, the southeastern states are dominated by southerly
winds in the summer. A marked change occurs in September, when a distinct anticyclonic circulation
appears in the region, persisting into December.
Not surprisingly, the more rugged topography at stations in the west produces a high degree of
spatial variability in mean wind direction (Figure 4). It is difficult to identify seasonal directional shifts
that are coherent throughout the region: mean wind directions at many stations are remarkably
consistent from one month to the next (Figure 5). Perhaps a more unexpected finding is the area of
very consistent wind directions extending from the eastern Great Lakes into New England. This axis
may be indicative of the climatological mean position of the ‘contact zone’ between continental and
maritime air masses in the eastern United States.
4.4. Discussion: surface winds, sea le6el pressure and air masses
In general, the strongest wind velocities east of the Rockies occur in winter for the northern United
States, in summer for the south-central states and southern Florida, and in spring and autumn for the
southeastern United States. In the west, broad regional patterns are less obvious owing to the effect
of topography on the local wind field. Seasonal cycles are most evident along the Front Range of
the Rockies, where velocities are consistently strong from late autumn to early spring, and along the
west coast of the United States, where velocities are highest from late spring to early autumn. The
strongest winds along the Front Range roughly coincide with the establishment of a high pressure
area over the Great Basin, which first appears in October and persists into March (Figure 6). A
northeastward-trending pressure gradient along the east coast of the United States also exists, with
higher pressure in the southeast (the southeast anticyclone; Bryson and Hare, 1974) and lower pressure
toward Maine. Wind directions from the Upper Midwest and into the northeast reflect this alignment
of isobars.
Copyright © 1999 Royal Meteorological Society
Int. J. Climatol. 19: 471 – 488 (1999)
478
Copyright © 1999 Royal Meteorological Society
K. KLINK
Int. J. Climatol. 19: 471 – 488 (1999)
Figure 3. Mean monthly wind speed, 1961 – 1990. The data are classified into four categories: B 3 m s − 1 (smallest dots); 3–4 m s − 1; 4–5 m s − 1; and \ 5 ms − 1 (largest dots),
and are plotted at station locations
479
Int. J. Climatol. 19: 471 – 488 (1999)
VARIANCE OF UNITED STATES SURFACE WIND SPEED, DIRECTION AND VELOCITY
Copyright © 1999 Royal Meteorological Society
Figure 4. Mean monthly wind direction, 1961 – 1990. The arrows (unit magnitude) are centered on station locations and fly with the wind
480
K. KLINK
Figure 5. Mean monthly wind direction for all months
By April, as the thermal low develops in the southwestern United States and the Bermuda/Azores high
builds into the southeast, velocities in the north-central and Great Lakes states begin to diminish, while
those in the southwestern and south-central states start to increase and to take on a southerly component
(somewhat modified by topography in the southwest). The Bermuda/Azores high remains dominant in the
east from May to September, and most stations east of the Rockies have predominantly southerly winds.
West coast stations exhibit their strongest wind velocities from April to August. The well-developed
thermal low enhances velocities in the southwest, while a consistent sea breeze contributes to increased
velocities along the west coast.
The velocity field reflects the trajectories of common air masses, particularly east of the Rockies. In the
northern states, Arctic and Pacific air is common in late autumn, throughout the winter and into spring;
Atlantic air masses are prevalent in late spring, summer and into the autumn (Bryson and Hare, 1974).
In the southern states, surface winds are dominated either by the southeastern anticyclone or by the
Bermuda/Azores high. Surface flow in the winter months is also reminiscent of the typical tracks of
cyclones and anticyclones over the United States (Zishka and Smith, 1980; Whittaker and Horn, 1984;
Harman, 1987). In the summer, the most common cyclone and anticyclone paths are close to or north of
the United States’ border, and their correspondence to the mean velocity maps is less clear. Overall, the
mean wind speed, direction and velocity patterns described here are very similar to those found in the
Climatic Atlas of the United States (U.S. Department of Commerce, 1968), suggesting that these surface
patterns are robust features of the mean wind field.
5. INTERANNUAL VARIABILITY OF THE SURFACE WIND FIELD
Wind speed, direction, and velocity variances are computed from the individual monthly means for the
period 1961–1990. The variances presented here therefore represent the year-to-year variability of the
mean wind field for a given month, rather than the wind variability typically experienced within a
particular month.
Copyright © 1999 Royal Meteorological Society
Int. J. Climatol. 19: 471 – 488 (1999)
481
Int. J. Climatol. 19: 471 – 488 (1999)
VARIANCE OF UNITED STATES SURFACE WIND SPEED, DIRECTION AND VELOCITY
Copyright © 1999 Royal Meteorological Society
Figure 6. Mean monthly sea level pressure, 1961 – 1990. The isobar interval is 2 hPa. Monthly mean sea level pressure was computed from sea level-corrected hourly and
three-hourly station pressures. Pressure corrections included dependence on temperature and humidity
482
K. KLINK
5.1. Velocity 6ariance
Wind velocity variability is related both to wind speed and wind direction variance. In general, the
velocity variance is not the same as the sum of the speed and direction variances (cf. Klink, 1998),
although stations with high variance in speed or direction tend also to have high velocity variance.
Overall, the wind velocity variance is highest in winter and spring (Figure 7), the time of year with the
most frequent occurrence of midlatitude cyclones and anticyclones (Zishka and Smith, 1980; Whittaker
and Horn, 1984; Harman, 1987). Change in the year-to-year frequency of synoptic-scale weather systems
(Zishka and Smith, 1980; Harman, 1987) is one possible contributor to the high velocity variance at these
times of the year. Summer and autumn velocity variance is low for most of the country. Cyclone and
anticyclone passages are less frequent during these seasons, so that semipermanent pressure patterns—the
thermal low in the southwest and the Bermuda/Azores high in the east—have a larger effect on velocity
variance. Low variability is also partly a function of the lower summer vector speeds. May shows a
distinct area of high variance from Texas to the Canadian border, which appears to coincide with the
onset of the ‘summer mode’ of the Bermuda/Azores high (Davis et al., 1997). For any given month, the
southeastern United States has among the lowest velocity variance in the country. The persistence of the
southeast anticyclone (Figure 6), with its typically lower wind speeds (Figure 3), contributes to the low
velocity variance.
5.2. Speed 6ariance
In general, wind speed variance is higher during the winter than during the summer (Figure 8), although
many stations show moderate levels of variance throughout the year. Lower interannual variance in the
summer is partly a function of the lower mean wind speeds. Speed variance increases from late autumn
into early spring, when the jet stream is positioned over the country and synoptic-scale weather systems
become more frequent.
Geographical patterns in speed variability are less distinct than those that appear in the velocity
variance fields. Locations along the Front Range of the Rockies show some of the highest speed variances
in a given month, perhaps partly related to their high mean wind speeds (Figure 3). Variance is generally
low in the southeastern United States, as are the mean speeds. Quillayute, Washington, is unusual in its
consistently high speed variance throughout the year. Quillayute is one of the stations with an incomplete
station history (Table I), and its 30-year time series of mean monthly wind speed suggests that, prior to
August 1966, its anemometer was positioned high above ground level (Figure 9). Faster wind speeds early
in the record are particularly evident in January, but July wind speeds, although smaller, show the same
trend. The higher speeds early in the station record appear to have biased the calculation of interannual
variance.
5.3. Direction 6ariance
Directional variability has a distinct seasonal cycle in the eastern United States, with higher variance
from late winter into spring and lower values during the summer (Figure 10). In the winter, directional
variance is highest in the south of the country, and the area of highest variance moves northward as
spring approaches. Variability diminishes during the summer and increases again in the autumn, first in
the north and later towards the south. One exception to this pattern appears in the southeastern United
States, where directional variance is high in nearly all months. Directional variance is low in most months
for the eastern Great Lakes and New England stations, as expected given the consistency of wind
direction (Figure 5). Seasonal changes in direction variance mirror the annual variation in mean sea level
pressure, stations with high variance typically being associated with centres of high pressure and/or with
weak pressure gradients (Figure 6). The high variance throughout the year in the southeast is associated
with both the Bermuda/Azores high and the southeastern anticyclone, and the low variance in the eastern
Great Lakes and New England may be related to a consistent mean location for the boundary between
continental and maritime air masses.
Copyright © 1999 Royal Meteorological Society
Int. J. Climatol. 19: 471 – 488 (1999)
483
Int. J. Climatol. 19: 471 – 488 (1999)
VARIANCE OF UNITED STATES SURFACE WIND SPEED, DIRECTION AND VELOCITY
Copyright © 1999 Royal Meteorological Society
Figure 7. Interannual wind velocity variance, 1961 – 1990. The data are classified into four categories: B 0.8 m2 s − 2 (smallest dots); 0.8–1.0 m s − 2; 1.0–1.2 m2 s − 2; and \ 1.2
m2 s − 2 (largest dots), and are plotted at station locations
484
Copyright © 1999 Royal Meteorological Society
K. KLINK
Int. J. Climatol. 19: 471 – 488 (1999)
Figure 8. Interannual wind speed variance, 1961 – 1990. The data are classified into four categories: B 0.1 m2 s − 2 (smallest dots); 0.1–0.3 m2 s − 2; 0.3–0.5 m2 s − 2; and \ 0.5
m2 s − 2 (largest dots), and are plotted at station locations
VARIANCE OF UNITED STATES SURFACE WIND SPEED, DIRECTION AND VELOCITY
485
In the west, a seasonal cycle in direction variance is less distinct. Many stations show their highest and
lowest values at different times of the year. Along the Front Range of the Rockies there is low variance
from autumn into spring and higher values during the summer, while along the west coast the directional
variance is higher in the winter and low in the summer (Figure 10). Low variance along the Front Range
approximately coincides with the establishment of cold-season high pressure over the west (Figure 6),
when fast winds and topographic channeling are reinforced by the surface pressure gradient. When the
high disappears in summer, directional variance increases. Along the west coast, onshore (sea breeze) wind
directions are quite consistent in the summer. During the winter, wind directions become more variable
as the land–sea contrast diminishes.
5.4. Summary: comparing 6elocity, speed and direction 6ariances
Patterns of variance in wind velocity share some characteristics with those of wind speed and direction.
In general, all three variances are larger in winter and spring and are smaller during the summer. Areas
with high variability in speed and direction also have high velocity variance (e.g. the Great Lakes states
in February). High variance can also result from high values in speed or direction separately; increased
velocity variance in winter along the Front Range, or in Texas during the spring, results mainly from high
speed variability, whereas high direction variability is an important contributor to the high winter velocity
variance in southern Texas and southern Florida. In the southeastern United States, high directional
variance and moderate speed variability have not yielded comparably high velocity variance, because the
low mean velocities (Figure 2) keep the velocity variance small.
6. SUMMARY AND CONCLUSIONS
Although each presents a different type of information, maps of the monthly mean wind speed, direction
and velocity share many similarities. Mean fields reflect the changing sea level pressure pattern over the
coterminous United States, particularly in the east where there is relatively little topographic variation. In
general, winter winds are predominantly from the west and northwest, while summer winds are primarily
Figure 9. The 1961–1990 time series of mean monthly wind speed (m s − 1) at Quillayute, Washington: (a) January; (b) July
Copyright © 1999 Royal Meteorological Society
Int. J. Climatol. 19: 471 – 488 (1999)
486
Copyright © 1999 Royal Meteorological Society
K. KLINK
Int. J. Climatol. 19: 471 – 488 (1999)
Figure 10. Interannual wind direction variance, 1961 – 1990. The data are classified into four (dimensionless) categories: B 0.3 (smallest dots); 0.3–0.5; 0.5–0.7; and \0.7 (largest
dots), and are plotted at station locations
VARIANCE OF UNITED STATES SURFACE WIND SPEED, DIRECTION AND VELOCITY
487
from the south; directions are more variable during the spring and autumn transition seasons. Wind
speeds (and vector speeds) are generally highest in winter and spring when the equator-to-pole temperature and pressure gradients are largest. In the west of the country, local effects on surface winds can be
as important as synoptic-scale controls. During the winter, high pressure over the Great Basin contributes
to high scalar and vector wind speeds along the Front Range, while directions are largely governed by the
local topography. Along the west coast during the summer, the enhanced land–sea contrast yields winds
that are strong and consistent in both speed and direction. During all seasons, wind directions and
velocities at some stations in the intermountain west are nearly identical from month to month.
Variance fields also show a distinct seasonality. Variance is generally higher in winter and spring, the
times of year with the most frequent occurrence of cyclones and anticyclones. Variance is often higher in
the east than in the west, because in the west, local features can restrict the range of surface wind patterns.
Velocity variance is a function of both speed and direction variance, as well as of the mean wind fields.
Velocity variance is generally high when either speed or direction variance is high. Low velocity variance
can result from low speed and direction variance, or from low vector wind speeds.
Similarities and differences in the mean wind speed, direction, and velocity fields and their interannual
variability result from the interaction of synoptic- and local-scale controls on the surface wind field.
Evaluating the mean and variance maps together aids in identifying features (such as mean sea level
pressure or surface topography) that are likely to have produced the observed patterns. Further work is
required to establish more firmly the relationship between synoptic-scale features and surface winds, and
the manner in which this relationship is modified in the presence of strong surface controls. Nevertheless,
the mean and variance climatologies presented here can serve as useful adjuncts to existing climatologies
of surface temperature and precipitation, and can enhance our understanding of the surface climatology
of the coterminous United States.
ACKNOWLEDGEMENTS
I thank the Minnesota Climatology Working Group for helpful discussion during the early stages of this
work, and S.M. Robeson and R.H. Skaggs for thoughtful comments on the manuscript. M.B. Lindberg
and the University of Minnesota Cartography Laboratory prepared the final figures. Portions of this
research were supported by the Minnesota Supercomputer Institute.
REFERENCES
Batschelet, E. 1981. Circular Statistics in Biology, Academic Press, New York.
Breckling, J. 1989. ‘The analysis of directional time series: Applications to wind speed and direction’. Lecture Notes in Statistics, Vol.
61, Springer, Berlin.
Bryson, R.A. and Hare, F.K. 1974. ‘The climates of North America’, in Bryson, R.A. and Hare, F.K. (eds), World Sur6ey of
Climatology, Vol. 11: Climates of North America, Elsevier, Amsterdam, pp. 1 – 47.
Burt, P.J.A. and Pedgley, D.E. 1997. ‘Nocturnal insect migration: effects of local winds’, Ad6. Ecol. Res., 27, 61 – 92.
Cabezudo, B., Recio, M., Sánchez-Laulhé, J., Del Mar Trigo, M., Toro, F.J., and Polvorinos, F. 1997. ‘Atmospheric transportation
of marihuana pollen from North Africa to the southwest of Europe’, Atmos. En6., 31, 3323 – 3328.
Changery, M.J. 1978. National Wind Data Index Final Report, U.S. Department of Energy Report HCO/T1041-01. Asheville, NC:
National Climatic Data Center.
Davis, R.E., Hayden, B.P., Gay, D.A., Phillips, W.L. and Jones, G.V. 1997. ‘The North Atlantic subtropical anticyclone’, J. Clim.,
10, 728 – 744.
Elliott, D.L., Holladay, C.G., Barchet, W.R., Foote, H.P. and Sandusky, W.F. 1986. Wind Energy Resource Atlas of the United
States, U.S. Department of Energy Report DOE/CH 10093-4. Golden, CO: Solar Energy Research Institute.
Feliks, Y., Gavze, E. and Givati, R. 1996. ‘Optimal vector interpolation of wind fields’, J. Appl. Meteorol., 35, 1153 – 1165.
Fisher, N.I. 1987. ‘Problems with the current definitions of the standard deviation of wind direction’, J. Clim. Appl. Meteorol., 26,
1522 – 1529.
Fisher, N.I. 1993. Statistical Analysis of Circular Data, Cambridge University Press, Cambridge.
Gaile, G.L. and Burt, J.E. 1980. Directional Statistics. Concepts and Techniques in Modern Geography, No. 25, Geo Books, Norwich,
CT.
Hanson, B., Klink, K., Matsuura, K., Robeson, S.M. and Willmott, C.J. 1992. ‘Vector correlation: review, exposition, and
geographic application’, Ann. Assoc. Am. Geogr., 82, 103 – 16.
Harman, J.R. 1987. ‘Mean monthly North American anticyclone frequencies, 1950 – 79’, Mon. Weather Re6., 115, 2840 – 2848.
Holton, J.R. 1979. An Introduction to Dynamic Meteorology, 2nd edn, Academic Press, New York.
Copyright © 1999 Royal Meteorological Society
Int. J. Climatol. 19: 471 – 488 (1999)
488
K. KLINK
Jönsson, P. and Fortuniak, K. 1995. ‘Interdecadal variations of surface wind direction in Lund, southern Sweden, 1741 – 1990’, Int.
J. Climatol., 15, 447–461.
Klink, K. 1998. ‘Complementary use of scalar, directional, and vector statistics with an application to surface winds’, Prof. Geogr.,
50, 3 – 13.
Klink, K. and Willmott, C.J. 1989. ‘Principal components of the surface wind field in the United States: a comparison of analyses
based upon wind velocity, direction, and speed’, Int. J. Climatol., 9, 293 – 308.
Legates, D.R. and Willmott, C.J. 1990. ‘Mean seasonal and spatial variability in gauge-corrected, global precipitation’, Int. J.
Climatol., 10, 111 –127.
Legler, D.M. 1983. ‘Empirical orthogonal function analysis of wind vectors over the tropical Pacific region’, Bull. Am. Meteorol.
Soc., 64, 234 – 241.
Mardia, K.V. 1972. Statistics of Directional Data, Academic Press, New York.
Martner, B.E. and Marwitz, J.D. 1982. ‘Wind characteristics in southern Wyoming’, J. Appl. Meteorol., 21, 1815 – 827.
NOAA (National Oceanic and Atmospheric Administration). 1993. Local Climatological Data – Annual Summaries for 1993, Vols
1 – 5, National Climatic Data Center, Asheville, NC.
NREL (National Renewable Energy Laboratory). 1993. Solar and Meteorological Surface Obser6ation Network 1961 – 1990, Ver. 1.0
(three volume CD set), National Climatic Data Center, Asheville, NC.
Oehlert, G.W. 1983. ‘Some statistics for wind roses’, Atmos. En6., 17, 2473 – 2476.
Palutikof, J.P., Kelly, P.M., Davies, T.D. and Halliday, J.A. 1987. ‘Impacts of spatial and temporal windspeed variability on wind
energy output’, J. Clim. Appl. Meteorol., 26, 1124–1133.
Peterson, E.W. and Hennessey, J.P., Jr. 1978. ‘On the use of power laws for estimates of wind power potential’, J. Appl. Meteorol.,
17, 390 – 394.
Roads, J.O., Chen, S.-C. and Ueyoshi, K. 1995. ‘Comparison of NMC’s global pressure analysis to NCDC’s U.S. observations’, J.
Clim., 8, 1410 – 1428.
Turner, D.B. 1986. ‘Comparison of three methods for calculating the standard deviation of the wind direction’, J. Clim. Appl.
Meteorol., 25, 703 –707.
U.S. Department of Commerce. 1968. Climatic Atlas of the United States, Environmental Science Services Administration/National
Oceanic and Atmospheric Administration, Washington, DC.
USGS (United States Geological Survey). 1970. The National Atlas of the United States of America, United States Geological Survey,
Washington, DC.
Visher, S.S. 1954. Climatic Atlas of the United States, Harvard University Press, Cambridge, MA.
Weber, R.O. 1997. ‘Estimators for the standard deviation of horizontal wind direction’, J. Appl. Meteorol., 36, 1403 – 1415.
Whittaker, L.M. and Horn, L.H. 1984. ‘Northern Hemisphere extratropical cyclone activity for four mid-season months’, J.
Climatol., 4, 297 – 310.
Zishka, K.M. and Smith, P.J. 1980. ‘The climatology of cyclones and anticyclones over North America and surrounding ocean
environs for January and July, 1950–77’, Mon. Weather Re6., 108, 387 – 401.
Copyright © 1999 Royal Meteorological Society
Int. J. Climatol. 19: 471 – 488 (1999)
Документ
Категория
Без категории
Просмотров
4
Размер файла
2 549 Кб
Теги
990
1/--страниц
Пожаловаться на содержимое документа