• Bove, M. C., 1998: Impacts of ENSO on United States tornadic activity. Preprints, Ninth Symp. on Global Change Studies, Phoenix, AZ, Amer. Meteor. Soc., 199–202.

    • Search Google Scholar
    • Export Citation
  • Bove, M. C., J. B. Elsner, C. W. Landsea, X. Niu, and J. J. O'Brien, 1998: Effect of El Niño on U.S. landfalling hurricanes, revisited. Bull. Amer. Meteor. Soc., 79 , 24772482.

    • Search Google Scholar
    • Export Citation
  • Gershunov, A., and T. P. Barnett, 1998a: ENSO influence on intraseasonal extreme rainfall and temperature frequencies in the contiguous United States: Observations and model results. J. Climate, 11 , 15751586.

    • Search Google Scholar
    • Export Citation
  • Gershunov, A., and T. P. Barnett, 1998b: Interdecadal modulation of ENSO teleconnections. Bull. Amer. Meteor. Soc., 79 , 27152725.

  • Glantz, M. H., 1988: Seasonal Responses to Regional Climatic Change: Forecasting by Analogy. Westview Press, 428 pp.

  • Green, P. M., 1996: Regional analysis of Canadian, Alaskan, and Mexican precipitation and temperature anomalies for ENSO impact. Center for Ocean–Atmospheric Prediction Studies, The Florida State University, Tech. Rep. 96-6, 104 pp. [Available from COAPS, The Florida State University, Tallahassee, FL 32306-2840.].

    • Search Google Scholar
    • Export Citation
  • Groisman, P. Ya, and H. P. Barker, 2002: Homogeneous blended wind data over the contiguous United States. Preprints, 13th Conf. on Applied Climatology, Portland, OR, Amer. Meteor. Soc., 114–117.

    • Search Google Scholar
    • Export Citation
  • Hanley, D. E., M. A. Bourassa, J. J. O'Brien, S. R. Smith, and H. M. Spade, 2003: A quantitative evaluation of ENSO indices. J. Climate, 16 , 12491258.

    • Search Google Scholar
    • Export Citation
  • Hollander, M., and W. A. Wolfe, 1999: Nonparametric Statistical Methods. John Wiley and Sons, 779 pp.

  • Horel, J. D., and J. M. Wallace, 1981: Planetary-scale atmospheric phenomena associated with the Southern Oscillation. Mon. Wea. Rev., 109 , 813829.

    • Search Google Scholar
    • Export Citation
  • Hurrell, J. W., 1996: Influence of variations in extratropical wintertime teleconnections on Northern Hemisphere temperature. Geophys. Res. Lett., 23 , 665668.

    • Search Google Scholar
    • Export Citation
  • Japan Meteorological Agency, 1991: Climate charts of sea surface temperatures of the western North Pacific and the global ocean. Marine Department, Japanese Meteorological Agency, 51 pp.

    • Search Google Scholar
    • Export Citation
  • Ludlum, D., 1991: The Audubon Society Field Guide to North American Weather. Alfred A. Knopf, 656 pp.

  • NCDC, 1998: Cooperative station summary of day data for the U.S. National Climatic Data Center Digital Dataset TD3200, 22 pp.

  • NCDC, 2001: Enhanced hourly wind station data for the contiguous United States. National Climatic Data Center Digital Dataset TD6421, 11 pp.

    • Search Google Scholar
    • Export Citation
  • Philander, S. G., 1990: El Niño, La Niña, and the Southern Oscillation. Academic Press, 293 pp.

  • Ropelewski, C. F., and M. S. Halpert, 1986: North American precipitation and temperature patterns associated with the El Niño–Southern Oscillation (ENSO). Mon. Wea. Rev., 114 , 23522362.

    • Search Google Scholar
    • Export Citation
  • Ropelewski, C. F., and M. S. Halpert, 1996: Quantifying Southern Oscillation precipitation relationships. J. Climate, 9 , 10431059.

  • Sevruk, B., 1982: Methods of correction for systematic error in point precipitation measurement for operational use. WMO Operational Hydrology Tech. Rep. 21, 91 pp.

    • Search Google Scholar
    • Export Citation
  • Sittel, M. C., 1994a: Marginal probabilities of the extremes of ENSO events for temperature and precipitation in the southeastern United States. COAPS Tech. Rep. 94-1, 153 pp. [Available from COAPS, The Florida State University, Tallahassee, FL 32306-2840.].

    • Search Google Scholar
    • Export Citation
  • Sittel, M. C., 1994b: Differences in the means of ENSO extremes for maximum temperature and precipitation in the United States. COAPS Tech. Rep. 94-2, 44 pp. [Available from COAPS, The Florida State University, Tallahassee, FL 32306-2840.].

    • Search Google Scholar
    • Export Citation
  • Smith, S. R., and J. J. O'Brien, 2001: Regional snowfall distributions associated with ENSO: Implications for seasonal forecasting. Bull. Amer. Meteor. Soc., 82 , 11791191.

    • Search Google Scholar
    • Export Citation
  • Smith, S. R., P. M. Green, A. P. Leonardi, and J. J. O'Brien, 1998: Role of multiple level tropospheric circulations in forcing ENSO winter precipitation anomalies. Mon. Wea. Rev., 126 , 31023116.

    • Search Google Scholar
    • Export Citation
  • Uccellini, L. W., 1990: Process contributing to the rapid development of extratropical cyclones. Extratropical Cyclones: The Erik Palmén Memorial Volume, C. W. Newton and E. O. Holopainen, Eds., Amer. Meteor. Soc., 81–105.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., and D. S. Gutzler, 1981: Teleconnections in the geopotential height field during the Northern Hemisphere winter. Mon. Wea. Rev., 109 , 784811.

    • Search Google Scholar
    • Export Citation
  • Yarnal, B., and H. F. Diaz, 1986: Relationships between extremes of the Southern Oscillation and the winter climate of the Anglo-American Pacific coast. J. Climatol., 6 , 197219.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    The map displays the percent change in monthly mean peak wind gust for the La Niña phase relative to neutral phase for (a) Nov, (b) Dec, (c) Jan, (d) Feb, and (e) Mar. All stations plotted meet the completeness requirements stated (see text sections 2 and 3). The magnitude and sign of the mean difference are represented by the size and shape of the symbols, respectively. Filled (hollow) symbols represent stations whose La Niña and neutral distributions are rejected (not rejected) as being equal at the 5% level of significance, according to the K–S test

  • View in gallery

    The map displays the difference in frequency of the number of days in the month that experience a minimum 28 kt (14.4 m s−1) wind gust for the La Niña phase relative to neutral phase for (a) Nov, (b) Dec, (c) Jan, (d) Feb, and (e) Mar. All stations plotted meet the completeness requirements. The magnitude and sign of the mean difference are represented by the size and shape of the symbols, respectively

  • View in gallery

    (a) Percentage of stations studied with significantly different warm and neutral phase distributions (gray) and cold and neutral phase distributions (black), as determined by the K–S test, where the dashed line indicates the critical value above which the chances of the probability of the given percentage of stations having significantly different distributions occurring by random chance is <1%, and (b) percentage of stations studied with a greater than 10% change in the monthly mean peak wind gust relative to neutral for the El Niño (gray) and La Niña (black)

  • View in gallery

    Same as Fig. 1, except for El Niño phase relative to neutral phase for (a) Oct, (b) Nov, (c) Jan, (d) Feb, (e) Apr, and (f) Jun

  • View in gallery

    Same as Fig. 2, except for El Niño phase relative to neutral phase for (a) Oct, (b) Nov, (c) Jan, (d) Feb, (e) Apr, and (f) Jun

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 203 178 3
PDF Downloads 54 43 2

ENSO Impacts on Peak Wind Gusts in the United States

View More View Less
  • 1 National Climatic Data Center, Asheville, North Carolina
  • | 2 Center for Ocean–Atmospheric Prediction Studies, The Florida State University, Tallahassee, Florida
© Get Permissions
Full access

Abstract

Changes in the peak wind gust magnitude in association with the warm and cold phases of the El Niño–Southern Oscillation (ENSO) are identified over the contiguous United States. All calculations of the peak wind gust are differences in the extreme phases of ENSO (warm and cold) relative to neutral for all stations in the study that pass the completeness criteria. Monthly composites were created for all years in the study (1 January 1948 through 31 August 1998). The differences in the mean peak wind gust are calculated for each month. A nonparametric statistical test was invoked to determine significant shifts in the extreme phase distributions. Differences in the frequency of gale-force wind gusts were also calculated.

The results show a dominant, ENSO cold phase wintertime signal. Regions most greatly affected are the Pacific Northwest, Southwest, the Great Plains, and the region extending from the Great Lakes through the Ohio River valley, and southwest into Texas. During the cold phase months from November to March, these regions experience an overall increase in the gustiness of the winds. The warm phase is associated with overall decreased gustiness in the Pacific Northwest during these months; however, the signal is of a lesser magnitude. There is also an observed decrease in the central Great Plains during the warm phase months of April and June. These results, along with improved ENSO forecasting, can work toward mitigating adverse effects of strong wind gusts and increase the utilization of wind power.

Corresponding author address: Mr. Jesse Enloe, National Climatic Data Center, 151 Patton Ave, Rm 514, Asheville, NC 28801. Email: jesse.enloe@noaa.gov

Abstract

Changes in the peak wind gust magnitude in association with the warm and cold phases of the El Niño–Southern Oscillation (ENSO) are identified over the contiguous United States. All calculations of the peak wind gust are differences in the extreme phases of ENSO (warm and cold) relative to neutral for all stations in the study that pass the completeness criteria. Monthly composites were created for all years in the study (1 January 1948 through 31 August 1998). The differences in the mean peak wind gust are calculated for each month. A nonparametric statistical test was invoked to determine significant shifts in the extreme phase distributions. Differences in the frequency of gale-force wind gusts were also calculated.

The results show a dominant, ENSO cold phase wintertime signal. Regions most greatly affected are the Pacific Northwest, Southwest, the Great Plains, and the region extending from the Great Lakes through the Ohio River valley, and southwest into Texas. During the cold phase months from November to March, these regions experience an overall increase in the gustiness of the winds. The warm phase is associated with overall decreased gustiness in the Pacific Northwest during these months; however, the signal is of a lesser magnitude. There is also an observed decrease in the central Great Plains during the warm phase months of April and June. These results, along with improved ENSO forecasting, can work toward mitigating adverse effects of strong wind gusts and increase the utilization of wind power.

Corresponding author address: Mr. Jesse Enloe, National Climatic Data Center, 151 Patton Ave, Rm 514, Asheville, NC 28801. Email: jesse.enloe@noaa.gov

1. Introduction

The present study is motivated by a desire to better understand the impact of ENSO on the climate in the United States. The El Niño–Southern Oscillation (ENSO) cycle has been shown to have significant impacts on various atmospheric parameters and phenomena over the continental United States. Past studies have shown the association of ENSO with temperature (Ropelewski and Halpert 1986; Sittel 1994b), precipitation (Ropelewski and Halpert 1996; Smith et al. 1998), snowfall (Smith and O'Brien 2001), hurricane landfalls (Bove et al. 1998; Tartaglione et al. 2002, manuscript submitted to J. Climate), and even tornadic activity (Bove 1998) over the United States. The present study focuses on wind; more specifically, on the upper tail of the wind distribution, the peak wind gust magnitude, to identify patterns associated with the ENSO cold and warm phase cycles.

Wind gusts may have a wide range of influence on human activities. Structural engineers must have detailed knowledge of wind potential, because the force on a structure is proportional to the square of the wind speed. The occurrence of multiple high wind gusts has a much different impact on bridges, buildings, or other structures than does a constant wind (average). Knowledge of wind gust magnitudes and frequencies is critical in the construction of various structures. The potential magnitude of the wind gust is critical to the transportation industry, as well. A strong burst of wind can prove fatal at take off or landing; therefore, airport operators must be aware of seasons when risks are increased. A sudden severe gust of wind (e.g., gale force) can overturn small sea craft, complicate operation of tractor trailers, motor homes, or other high-profile vehicles, even aid in the spread of forest fires. Recreational activities, such as hang gliding or skydiving, are also greatly impacted by gusts of wind. In addition to safety, foreknowledge of this parameter can also prove valuable to those who harness the energy created by wind.

This study identifies the relationships of ENSO with peak wind gusts over the contiguous United States. The study considers the change in the monthly mean peak wind gust and the frequency of gale-force wind gusts. The region of ENSO's greatest influence is the Pacific Northwest. The results, stated in section 4, show a dominant cold phase signal during the fall and winter months countered by a weaker, less persistent warm phase signal. This new information on the association between ENSO and the peak wind gust, coupled with the increased capacity for forecasting ENSO warm and cold phases may lead to improvements in forecasting, thereby mitigating financial impacts, increasing safety awareness, and potentially increasing the benefits of wind power.

2. Data

Daily peak wind gust magnitude data were obtained from the TD3200 First-Order Summary of the Day (FSOD) dataset (NCDC 1998) provided by the National Climatic Data Center. The FSOD peak wind gust magnitude is defined as the highest, 5-s time-averaged magnitude of wind speed recorded in knots by a station's anemometer for a given 24-h period. The FSOD data contain observations from stations worldwide collected by certified observers from the National Weather Service (NWS), U.S. Air Force (Air Weather Service), U.S. Navy (Navy Weather Service), and the Federal Aviation Administration (FAA). The present study focuses on the contiguous United States.

The TD3200 dataset spans the period from 1 January 1948 to 31 August 1998. Wind speeds are expressed in knots from January 1955 onward. Prior to 1955, the winds speeds are expressed in miles per hour, with the exception of Navy stations, which used knots for the entire period of record. All winds were converted to knots for the present study.

Anemometer elevation throughout the network of stations varies greatly over the period of record (NCDC 2001) and, therefore, the elevation homogenization of the near-surface wind is necessary before any climatological evaluation can be performed. In this process, a standard logarithmic near-surface profile for unobstructed wind movement is assumed (because most of the sites are airport or coastal locations):
i1520-0442-17-8-1728-e1
where Ua is the wind speed at the anemometer height, Hsnod is the snow depth, Ha is the anemometer height above the ground, z0 is the surface roughness (a function of the presence of snow cover at the site), and U10m is the speed at 10 m above the ground (Groisman and Barker 2002). The presence of snow and its depth are obtained from the FSOD daily data. Assuming the meteorological instruments are stationed over grass (of height ≈ 5–20 cm), the roughness parameter z0 = 3 cm. For uniformly distributed snow cover, z0 = 0.5 cm (Sevruk 1982).

The number of active stations also varies throughout the study period. Of the stations that exist for the entire period, most have some missing data. A 90% completeness criterion is invoked to ensure a level of quality of the monthly averages computed from the data. Any month with less than 90% of the data is not included in the climatological calculations for that month at that station. A total of 169 stations pass this criterion and are included in the present study.

The classification of ENSO events followed in the present study is defined by the Japan Meteorological Agency (JMA) sea surface temperature (SST) index. The JMA index is found to be more sensitive to La Niña events than all other indices (Hanley et al. 2003). The JMA SST index defines phases of ENSO based on sea surface temperature anomalies in the region from 4°N to 4°S and from 150° to 90°W. A warm (cold) phase is defined when the 5-month running mean of SST anomalies in the defined region is greater than 0.5°C (less than −0.5°C) for at least six consecutive months—otherwise, the ENSO phase is classified as neutral (Japan Meteorological Agency 1991). The event must begin before the start of the ENSO year (October) and include October, November, and December (Sittel 1994a).

Extremes in ENSO typically develop during summer, climax in the fall, and subside the following spring. Therefore, an ENSO year is defined as beginning in October of the onset year and continuing through September of the following year (Green 1996). For example, the 1997 warm event year begins in October 1997 and ends in September 1998. This year is used to highlight the effects of the ENSO event from its maturity in the fall.

3. Methodology

Monthly averages of the peak wind gust are computed for the 169 stations. The averages are then classified according to ENSO phase (Table 1). For the 51 yr of data, from 1 January 1948 to 31 August 1998, there are 12 warm phases (11 for September, because the dataset ended in August of the 1997 warm phase), 28 neutral phases (27 for October, November, and December, because the data do not begin until January of the 1947 neutral phase year), and 11 cold phases for each month.

Long-term monthly averages are computed for each phase. Calculating a station's long-term monthly mean for the warm (neutral, cold) phase requires a minimum of 3 (7, 3) months to have peak wind averages during a warm (neutral, cold) phase. For example, to compute the January warm phase long-term monthly average for Spokane, Washington, there would need to be at least three warm phase Januaries that passed the 90% criterion in the Spokane daily peak wind data. Calculations using higher thresholds (5 warm, 10 neutral, 5 cold; and 7 warm, 12 neutral, 7 cold) were also examined and revealed patterns similar to those observed using the lesser criteria.

Maps are generated to spatially plot differences in the ENSO extreme event (warm, cold) peak wind gusts relative to neutral (Figs. 1 and 4 later). The neutral phase is used as the base for this study, because mean values are influenced by the extremes. The percent change [Eq. (2)] is calculated for each month at each station:
i1520-0442-17-8-1728-e2

Stations that exhibit significant changes in peak wind during the extreme ENSO phases are determined by constructing monthly distributions for each phase at each station from the daily observations. The Kolmogorov–Smirnov (K–S) test is used to ascertain whether the peak wind gust distribution of the ENSO extreme event for a given month is significantly different from the corresponding neutral distribution. The K–S test is a distribution-free test for general differences in two populations and tests the difference of the entire distribution (Hollander and Wolfe 1999). The test does not reveal significant differences at any particular percentile. For the present study, the observed significance level at which the distributions are considered significantly different is 5%. For further detail on the K–S test, see the appendix. The results of the K–S test are plotted spatially along with the difference in means (Figs. 1 and 4).

In 108 independent experiments (the minimum number of stations that passed the completeness criteria for any given month/phase) with a chance to succeed equal to 5%, 10 experiments (9.26%) would be successful just by random chance, with a probability of 1%. By applying a field significance test it can be stated that if more than 9.26% of the stations for a given month and phase have significantly different distributions, the probability that the reported pattern is occurring by random chance is less than 1% (Fig. 3a).

The probability of a severe event (i.e., a high wind speed) is considered to quantify the upper tails of the distribution of peak wind gust. The authors consider the probability of a peak wind gust that meets or exceeds a threshold of 28 kt (14.4 m s−1) the minimum classification of a gale-force (moderate gale) wind speed according to the terrestrial-modified Beaufort wind scale (Ludlum 1991). A gale-force wind gust has potential adverse effects on industry, as well as recreation, for maritime or terrestrial activities. A monthly mean frequency of gale-force gusts was computed at each station and classified by ENSO phase. The differences between the extreme and neutral phases are plotted spatially (Figs. 2 and 5 later).

4. Results

The strongest, most persistent peak wind gust signals occur during the fall and winter for both ENSO extremes. It is not uncommon for patterns associated with ENSO to have their peak magnitudes in these colder months, because both the warm and cold phases reach maturity during the fall (Glantz 1988; Philander 1990). Other wintertime peaks in observed signals associated with ENSO over the contiguous United States have been noted in previous studies, such as temperature (Sittel 1994b; Green 1996) and precipitation (Sittel 1994b; Smith et al. 1998). The peak wind gust patterns associated with the warm phase are not equal and opposite to those observed with the cold phase. There is a clear, dominant, cold phase wintertime signal. Relative to neutral, the cold phase is characterized by more stations that demonstrate significant shifts in distribution, larger differences in the monthly mean, and a higher change in frequency of gale-force wind gusts than occur during a warm phase. The signals associated with El Niño lack consistency in temporal and spatial patterns (Figs. 4 and 5 both later). The region with the highest contrast in signals between the two extreme phases is the Northwest (Washington, Oregon, Idaho, Montana, Wyoming, and Colorado). The area is composed of large, positive shifts in the mean peak wind gusts and frequency of gale-force gusts during the cold phase winter months, while in the warm phase fall and winter months, smaller magnitudes of negative values dominate.

a. La Niña impact

Over most of the year and the majority of the country, La Niña is associated largely with positive differences in monthly means of the peak wind gust (Fig. 1) and the frequency of gale-force wind gusts (Fig. 2). One of the strongest, most persistent signals associated with the cold phase is focused about the Northwest in the fall and winter months (Fig. 1). The region over the Ohio River valley, from the Great Lakes, extending southwestward into Texas, is also observed to exhibit a cold phase signal during the period from October through March. There is an exception in late summer when the presence of these positive differences is not as evident (not shown). The regions that experience the strongest and most spatially consistent positive wind gust anomalies in January, February, and March (e.g., the Northwest and the Ohio River valley) have also been shown to experience positive rainfall anomalies during the same months (Gershunov and Barnett 1998a,b; Smith et al. 1998).

The observed signals associated with the cold phase of ENSO are characterized by larger magnitudes of differences in the monthly mean than are observed with warm phase signals, particularly from November to March. During an average, neutral November, Seattle, Washington, experiences a daily peak wind gust of 21.4 kt (11 m s−1). A cold phase November peak wind gust in Seattle averages 23.9 kt (12.3 m s−1), an 11.8% increase.

During the cold phase, the months from November to March display the highest percentage of stations that report a difference in means of greater than 10% (Fig. 3b). Consequently, the computation of the K–S test reveals these months to exhibit the highest percentage of stations whose cold phase and neutral phase distributions are significantly different (Fig. 3a). The number of stations for each month and phase is always higher than the critical value of 9.26% (see section 3), demonstrating that the reported patterns have a less than 1% chance of being random. The highest concentration of stations with these large differences in the mean and significant shifts in distribution during these months occur in the in the West, southern Great Plains, and the Ohio River valley. A signal, if any, could not be determined over the northern Great Plains due to the lack of data.

The patterns observed in this study experience a degree of month-to-month variability. During the La Niña January, positive differences over California are replaced by negative values along the Pacific Coast, but these differences are restricted to ∼42°N latitude, southward (Figs. 1c and 2c). Stronger, positive differences (greater than 20%) remain in the Pacific Northwest creating a north–south-oriented dipole pattern with values from −5% to −15% present along the Pacific Coast in California. The signal observed from the southern Great Plains through the Ohio River valley also subsides in this month (Figs. 1c and 2c). In the following two months, the strong, positive values return to the southern Great Plains, while the absence of these large differences in the Southwest remains (Figs. 1d–e, 2d–e).

Focusing on the number of days in a month where the peak wind gust meets or exceeds a 28-kt (14.4 m s−1) threshold, the difference in frequencies of such an occurrence between the neutral and extreme phases mimic the pattern observed by the difference in means of the monthly peak wind gust (Fig. 2). An average, neutral November in Seattle has 6.3 days that meet or exceed a peak wind gust of 28 kt (14.4 m s−1). However, during a cold phase November, Seattle will average 9.3 days that produce a minimum 28-kt (14.4 m s−1) wind gust.

b. El Niño impact

The observed patterns associated with the warm phase of ENSO are not as robust as signals observed during the cold phase. The differences in the warm and neutral monthly means are smaller and not as persistent from month to month. The surface observation stations record these percent changes as relatively weak, rarely greater than 10% (Fig. 4). Though strong differences in means are not observed, there is a general reduction in peak wind gust over the entire contiguous United States during El Niño, as well as a reduction in occurrences of gale-force wind gusts (Fig. 5). As in La Niña, this general effect is persistent throughout all seasons, with the exception of the July–August–September summer, when there is primarily a positive shift. Performing the K–S test reveals far fewer stations having significant differences in the warm and neutral phase distributions than occur between the cold and neutral phases (Fig. 3a).

One persistent warm phase signal, though weak, is observed in the Northwest (Figs. 4a–d). From the beginning of the El Niño year (October) through February, a general weak reduction in the monthly mean peak wind gust (usually 0% to −5%) spans most of the country. However, percent changes of −5% to −10% are more common in the Northwest, with some magnitudes as great as −15% to −20% occurring during these months. In a warm phase November, Seattle averages a daily peak wind gust of 10.3% smaller (19.2 kt, 9.9 m s−1) than the average neutral phase wind gust of 21.4 kt (11 m s−1). This is in direct contrast to the increase in the monthly mean peak wind gust observed during the cold phase. There is also a lower frequency of days that meet or exceed the 28-kt (14.4 m s−1) threshold in Seattle (Figs. 5a–d) with 4 days that experience a minimum 28-kt (14.4 m s−1) wind gust during a warm event, compared to the 6.3 days of the average neutral November. This pattern persists through February (Figs. 4a–d), with slight variation (lacking in the month of January).

The Great Plains also experiences a reduction in the mean peak wind gusts in the fall and winter months. The signal during these months in this region, however, is not as strong in magnitude or as persistent from month to month as what is observed in the Northwest. A reduction in peak wind gusts is also noted in the central Great Plains during the spring (Figs. 4e–f, 5e–f). This signal is most apparent as a reduced frequency in gale force gusts during April and June; however, the reduction is lacking in the month of May.

Though negative differences are present over most of the country for the month of October, weak, positive peak wind gust differences are present in the southwesternmost region of California. Through the winter, the positive values migrate northward through California. By January, the positive values extend to the Canadian border (possibly even farther north, but the data are limited to the United States) on the windward side of the Cascade Mountains (Fig. 4c). Negative differences (−5% to −15%) remain on the eastern, leeward side of the mountains, creating a weak, east–west dipole pattern in the Northwest during the month of January with small, positive values (mostly 5% to 10%) on the western, windward side of the Cascades. These weak, positive values are present in the Pacific Northwest during January, subside in February (Fig. 4d), giving way to negative differences in Oregon and Washington, and then return in March.

5. Discussion

In addition to the identification of wind gust signals associated with ENSO, it is also useful to consider the possible role that large-scale atmospheric circulations play in developing the observed wind gust patterns. Previous studies have established that ENSO is teleconnected to the upper-level flow over North America and, subsequently, to storm tracks (Horel and Wallace 1981; Wallace and Gutzler 1981; Smith et al. 1998). Linking the observed wind gust patterns to atmospheric dynamics can work toward improving ENSO-based forecasting. The authors limit the current discussion to proposing reasonable connections between physical mechanisms that could influence the peak wind gust in hopes of stimulating future discussion and research on the topic.

Precipitation can be used as an indication of the storminess in a region and has been shown to vary over the United States in association with ENSO phases. For example, the low-level onshore wind flow in the Pacific Northwest provides a moisture supply for precipitation from the Pacific Ocean and reaches a maximum during cold phase winters (Smith et al. 1998). Anomalously strong westerly flow of the upper-level jet and positive vorticity advection promote uplift, facilitating an anomalous increase in precipitation in the region (Smith et al. 1998; Yarnal and Diaz 1986).

During the cold phase winter months of January, February, and March, the regions that experience the strongest and most spatially consistent positive wind gust anomalies (Figs. 1 and 2; e.g., the Northwest and the Ohio River valley) also experience positive precipitation anomalies (Gershunov and Barnett 1998a,b; Sittel 1994b; Smith and O'Brien 2001; Smith et al. 1998). In contrast, the Northwest experiences an overall decrease in the gustiness of the winds and anomalously negative rainfall amounts (Figs. 4a–f, and 5a–f; Gershunov and Barnett 1998a; Smith et al. 1998) during warm phase winters.

In the La Niña January, a north–south dipole pattern forms in the Pacific Northwest (Figs. 1c–e and 2c–e), with increased (decreased) gustiness to the north (south). It should be noted that this dipole pattern is also seen in precipitation patterns during the La Niña winter (Sittel 1994b) with increased (decreased) precipitation to the north (south) of the California–Oregon border.

The present study also revealed a decrease in the gustiness of the wind in the Great Plains during the warm phase spring. Coherent with this observation is a decrease in the frequency of tornadoes in the Great Plains during the spring warm phase (Bove 1998), implying a decrease in tornadic storms—a potential source for strong wind gusts. Clearly, the current study and past research show a relationship between increased (or decreased) precipitation and more (or less) wind gustiness associated with ENSO phases.

Changes in the strength and position of an upper-level jet have been associated with ENSO precipitation anomalies (Smith et al. 1998). In a similar fashion, the authors propose that the upper jet can affect the gustiness of surface winds through the storms and cyclones that the jet generates. Uccellini (1990) has shown that ascending motion is associated with the entrance and exit regions of a core of stronger winds in the jet stream. The jet stream provides the upper-level dynamics necessary for positive vertical motion in the atmosphere, as well as momentum in the middle and upper levels. Regions on the south (north) side of the entrance (exit) of a zonally positioned jet in the Northern Hemisphere experience a large-scale ascending motion, while regions on the north (south) side of the exit (entrance) region experience a large-scale descending motion. This dynamic lift and the related synoptic-scale vorticity patterns are often associated with cyclonic activity. Cyclones that form are comprised of large-scale vertical motion. Downscaling (i.e., larger-scale dynamics successively interacting with smaller scales—global to synoptic, synoptic to mesoscale, etc.) from within the synoptic-scale system to individual, mesoscale convective events can result in the transfer of the jet stream momentum. The jet stream is three-dimensional and, therefore, its momentum is present from the upper to the middle levels of the troposphere. Convective cells within the large-scale cyclone whose tops reach these elevations can transfer this higher momentum through their downdrafts, bringing the colder, higher-velocity air to the surface. Through scale interaction, the dynamic lifting yields the potential for increased gustiness over the region. Further analysis will be required to confirm the proposed interaction between ENSO, jet streams, precipitation, and wind gustiness.

6. Conclusions

Shifts in the distribution of peak wind gust are identified in association with the warm and cold phases of ENSO. The significant shifts were determined using the Kolmogorov–Smirnov test, which also helped highlight the dominant La Niña wintertime signal (Fig. 3a). Monthly mean peak wind gusts, as well as the frequency of gale-force gusts, are found to increase (decrease) during ENSO cold (warm) phases over the West, particularly the Pacific Northwest, in the fall and winter. Gustiness is also observed to increase over the Ohio River valley, southwest into Texas, during the La Niña fall and winter. Decreased gustiness relative to neutral years is noted over the Great Plains during the warm phase. In many cases, incresases (decreases) in gustiness are associated with positive (negative) precipitation anomalies. Physical mechanisms connecting precipitation and wind gustiness may be related to a transfer of middle- and upper-tropospheric momentum to the surface; however, this connection must await future analysis.

Knowledge of the associated affects of ENSO on the magnitude of the wind gust and frequency of severe occurrences is essential for improvements in city planning, transportation, and recreation safety. Increasing public awareness can mitigate potential negative effects caused by severe wind gusts. The foreknowledge of ENSO impacts on wind may also help in attempts at harnessing the energy created by wind, taking advantage of the effects. The authors wish to note that ENSO is not the only global ocean–atmosphere phenomenon altering tropospheric circulation patterns. Other oscillations, such as the North Atlantic Oscillation (Hurrell 1996) and the Pacific decadal oscillation (Gershunov and Barnett 1998b), have also been found to impact flow patterns that affect weather globally and should be taken into consideration when developing forecasting methods.

Acknowledgments

This work is funded by the NOAA Office of Global Programs under the Applied Research Center at The Florida State University (Grant NA16GP1365), which supports NCEP and the IRI with improved knowledge of seasonal to interannual climate variability. NASA Grant GWEC-0000-0052 and the NOAA Climate and Global Change Program (Climate Change and Detection Element) provided partial support for this study.

Special thanks also to those individuals who helped make this research possible: Dr. Pavel Ya. Groisman at the National Climatic Data Center, Dr. Steven Morey at the Center for Ocean–Atmospheric Prediction Studies, and Peter Gavin for his programming assistance.

REFERENCES

  • Bove, M. C., 1998: Impacts of ENSO on United States tornadic activity. Preprints, Ninth Symp. on Global Change Studies, Phoenix, AZ, Amer. Meteor. Soc., 199–202.

    • Search Google Scholar
    • Export Citation
  • Bove, M. C., J. B. Elsner, C. W. Landsea, X. Niu, and J. J. O'Brien, 1998: Effect of El Niño on U.S. landfalling hurricanes, revisited. Bull. Amer. Meteor. Soc., 79 , 24772482.

    • Search Google Scholar
    • Export Citation
  • Gershunov, A., and T. P. Barnett, 1998a: ENSO influence on intraseasonal extreme rainfall and temperature frequencies in the contiguous United States: Observations and model results. J. Climate, 11 , 15751586.

    • Search Google Scholar
    • Export Citation
  • Gershunov, A., and T. P. Barnett, 1998b: Interdecadal modulation of ENSO teleconnections. Bull. Amer. Meteor. Soc., 79 , 27152725.

  • Glantz, M. H., 1988: Seasonal Responses to Regional Climatic Change: Forecasting by Analogy. Westview Press, 428 pp.

  • Green, P. M., 1996: Regional analysis of Canadian, Alaskan, and Mexican precipitation and temperature anomalies for ENSO impact. Center for Ocean–Atmospheric Prediction Studies, The Florida State University, Tech. Rep. 96-6, 104 pp. [Available from COAPS, The Florida State University, Tallahassee, FL 32306-2840.].

    • Search Google Scholar
    • Export Citation
  • Groisman, P. Ya, and H. P. Barker, 2002: Homogeneous blended wind data over the contiguous United States. Preprints, 13th Conf. on Applied Climatology, Portland, OR, Amer. Meteor. Soc., 114–117.

    • Search Google Scholar
    • Export Citation
  • Hanley, D. E., M. A. Bourassa, J. J. O'Brien, S. R. Smith, and H. M. Spade, 2003: A quantitative evaluation of ENSO indices. J. Climate, 16 , 12491258.

    • Search Google Scholar
    • Export Citation
  • Hollander, M., and W. A. Wolfe, 1999: Nonparametric Statistical Methods. John Wiley and Sons, 779 pp.

  • Horel, J. D., and J. M. Wallace, 1981: Planetary-scale atmospheric phenomena associated with the Southern Oscillation. Mon. Wea. Rev., 109 , 813829.

    • Search Google Scholar
    • Export Citation
  • Hurrell, J. W., 1996: Influence of variations in extratropical wintertime teleconnections on Northern Hemisphere temperature. Geophys. Res. Lett., 23 , 665668.

    • Search Google Scholar
    • Export Citation
  • Japan Meteorological Agency, 1991: Climate charts of sea surface temperatures of the western North Pacific and the global ocean. Marine Department, Japanese Meteorological Agency, 51 pp.

    • Search Google Scholar
    • Export Citation
  • Ludlum, D., 1991: The Audubon Society Field Guide to North American Weather. Alfred A. Knopf, 656 pp.

  • NCDC, 1998: Cooperative station summary of day data for the U.S. National Climatic Data Center Digital Dataset TD3200, 22 pp.

  • NCDC, 2001: Enhanced hourly wind station data for the contiguous United States. National Climatic Data Center Digital Dataset TD6421, 11 pp.

    • Search Google Scholar
    • Export Citation
  • Philander, S. G., 1990: El Niño, La Niña, and the Southern Oscillation. Academic Press, 293 pp.

  • Ropelewski, C. F., and M. S. Halpert, 1986: North American precipitation and temperature patterns associated with the El Niño–Southern Oscillation (ENSO). Mon. Wea. Rev., 114 , 23522362.

    • Search Google Scholar
    • Export Citation
  • Ropelewski, C. F., and M. S. Halpert, 1996: Quantifying Southern Oscillation precipitation relationships. J. Climate, 9 , 10431059.

  • Sevruk, B., 1982: Methods of correction for systematic error in point precipitation measurement for operational use. WMO Operational Hydrology Tech. Rep. 21, 91 pp.

    • Search Google Scholar
    • Export Citation
  • Sittel, M. C., 1994a: Marginal probabilities of the extremes of ENSO events for temperature and precipitation in the southeastern United States. COAPS Tech. Rep. 94-1, 153 pp. [Available from COAPS, The Florida State University, Tallahassee, FL 32306-2840.].

    • Search Google Scholar
    • Export Citation
  • Sittel, M. C., 1994b: Differences in the means of ENSO extremes for maximum temperature and precipitation in the United States. COAPS Tech. Rep. 94-2, 44 pp. [Available from COAPS, The Florida State University, Tallahassee, FL 32306-2840.].

    • Search Google Scholar
    • Export Citation
  • Smith, S. R., and J. J. O'Brien, 2001: Regional snowfall distributions associated with ENSO: Implications for seasonal forecasting. Bull. Amer. Meteor. Soc., 82 , 11791191.

    • Search Google Scholar
    • Export Citation
  • Smith, S. R., P. M. Green, A. P. Leonardi, and J. J. O'Brien, 1998: Role of multiple level tropospheric circulations in forcing ENSO winter precipitation anomalies. Mon. Wea. Rev., 126 , 31023116.

    • Search Google Scholar
    • Export Citation
  • Uccellini, L. W., 1990: Process contributing to the rapid development of extratropical cyclones. Extratropical Cyclones: The Erik Palmén Memorial Volume, C. W. Newton and E. O. Holopainen, Eds., Amer. Meteor. Soc., 81–105.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., and D. S. Gutzler, 1981: Teleconnections in the geopotential height field during the Northern Hemisphere winter. Mon. Wea. Rev., 109 , 784811.

    • Search Google Scholar
    • Export Citation
  • Yarnal, B., and H. F. Diaz, 1986: Relationships between extremes of the Southern Oscillation and the winter climate of the Anglo-American Pacific coast. J. Climatol., 6 , 197219.

    • Search Google Scholar
    • Export Citation

APPENDIX

Kolmogorov–Smirnov Test

Kolmogorov–Smirnov test is a distribution-free test for general differences in two populations (Hollander and Wolfe 1999).

Given two independent samples, X1, … , Xm and Y1, … , Ym, the null hypothesis H0 = the two distributions are equal is tested against the alternative H1 = they are different. Let Z1, … , ZN (where N = m + n) be the ordered values for the combined samples of X1, … , Xm and Y1, … , Ym. From this the empirical distribution functions A and B are obtained:
i1520-0442-17-8-1728-ea1a
where 1 ≤ iN. From this, the two-sided two-sample Kolmogorov–Smirnov test statistic J is computed:
i1520-0442-17-8-1728-ea2
If the observed significance level F(J) < α, H0 is rejected (this study uses the level of significance, α = 0.05), where
i1520-0442-17-8-1728-ea3
From (A3), if Jthe critical value = 1.3581, there is a probability of less than 5% that X1, … , Xm and Y1, … , Yn have equal distributions, and H0 is rejected in favor of H1.

Fig. 1.
Fig. 1.

The map displays the percent change in monthly mean peak wind gust for the La Niña phase relative to neutral phase for (a) Nov, (b) Dec, (c) Jan, (d) Feb, and (e) Mar. All stations plotted meet the completeness requirements stated (see text sections 2 and 3). The magnitude and sign of the mean difference are represented by the size and shape of the symbols, respectively. Filled (hollow) symbols represent stations whose La Niña and neutral distributions are rejected (not rejected) as being equal at the 5% level of significance, according to the K–S test

Citation: Journal of Climate 17, 8; 10.1175/1520-0442(2004)017<1728:EIOPWG>2.0.CO;2

Fig. 2.
Fig. 2.

The map displays the difference in frequency of the number of days in the month that experience a minimum 28 kt (14.4 m s−1) wind gust for the La Niña phase relative to neutral phase for (a) Nov, (b) Dec, (c) Jan, (d) Feb, and (e) Mar. All stations plotted meet the completeness requirements. The magnitude and sign of the mean difference are represented by the size and shape of the symbols, respectively

Citation: Journal of Climate 17, 8; 10.1175/1520-0442(2004)017<1728:EIOPWG>2.0.CO;2

Fig. 3.
Fig. 3.

(a) Percentage of stations studied with significantly different warm and neutral phase distributions (gray) and cold and neutral phase distributions (black), as determined by the K–S test, where the dashed line indicates the critical value above which the chances of the probability of the given percentage of stations having significantly different distributions occurring by random chance is <1%, and (b) percentage of stations studied with a greater than 10% change in the monthly mean peak wind gust relative to neutral for the El Niño (gray) and La Niña (black)

Citation: Journal of Climate 17, 8; 10.1175/1520-0442(2004)017<1728:EIOPWG>2.0.CO;2

Fig. 4.
Fig. 4.

Same as Fig. 1, except for El Niño phase relative to neutral phase for (a) Oct, (b) Nov, (c) Jan, (d) Feb, (e) Apr, and (f) Jun

Citation: Journal of Climate 17, 8; 10.1175/1520-0442(2004)017<1728:EIOPWG>2.0.CO;2

Fig. 5.
Fig. 5.

Same as Fig. 2, except for El Niño phase relative to neutral phase for (a) Oct, (b) Nov, (c) Jan, (d) Feb, (e) Apr, and (f) Jun

Citation: Journal of Climate 17, 8; 10.1175/1520-0442(2004)017<1728:EIOPWG>2.0.CO;2

Table 1.

ENSO phases (1947–97) based on the JMA SST index for the period of study. Each year indicates the beginning of the ENSO year (e.g., 1982 indicates a warm phase from Oct 1982 to Sep 1983)

Table 1.
Save