• Atlas, R., , R. N. Hoffman, , S. C. Bloom, , J. C. Jusem, , and J. Ardizzone, 1996: A multiyear global surface wind velocity dataset using SSM/I wind observations. Bull. Amer. Meteor. Soc., 77 , 869882.

    • Search Google Scholar
    • Export Citation
  • Bjerknes, J., 1969: Atmospheric teleconnections from the equatorial Pacific. Mon. Wea. Rev., 97 , 163172.

  • Chu, P-S., 1988: Extratropical forcing and the burst of equatorial westerlies in the western Pacific: A synoptic study. J. Meteor. Soc. Japan, 66 , 549564.

    • Search Google Scholar
    • Export Citation
  • Eisenman, I., , L. Yu, , and E. Tziperman, 2005: Westerly wind bursts: ENSO’s tail rather than the dog? J. Climate, 18 , 52245238.

  • Fedorov, A. V., 2002: The response of the coupled tropical ocean–atmosphere to westerly wind bursts. Quart. J. Roy. Meteor. Soc., 128 , 123.

    • Search Google Scholar
    • Export Citation
  • Fedorov, A. V., , S. L. Harper, , S. G. Philander, , B. Winter, , and A. Wittenberg, 2003: How predictable is El Niño? Bull. Amer. Meteor. Soc., 84 , 911920.

    • Search Google Scholar
    • Export Citation
  • Gill, A. E., 1980: Some simple solutions for heat-induced tropical circulation. Quart. J. Roy. Meteor. Soc., 106 , 447462.

  • Harrison, D. E., , and P. S. Schopf, 1984: Kelvin wave induced anomalous advection and the onset of surface warming in El Niño events. Mon. Wea. Rev., 112 , 923933.

    • Search Google Scholar
    • Export Citation
  • Harrison, D. E., , and B. S. Giese, 1991: Episodes of surface westerly winds as observed from islands in the western tropical Pacific. J. Geophys. Res., 96 , 32213237.

    • Search Google Scholar
    • Export Citation
  • Harrison, D. E., , and G. A. Vecchi, 1997: Westerly wind events in the tropical Pacific. J. Climate, 10 , 31313156.

  • Hartten, L. M., 1996: Synoptic settings of westerly wind bursts. J. Geophys. Res., 101 , 1699717019.

  • Hendon, H. H., , and M. L. Salby, 1994: The life cycle of the Madden–Julian oscillation. J. Atmos. Sci., 51 , 22252237.

  • Hendon, H. H., , B. Liebmann, , and J. D. Glick, 1998: Oceanic Kelvin waves and the Madden–Julian oscillation. J. Atmos. Sci., 55 , 88101.

    • Search Google Scholar
    • Export Citation
  • Hendon, H. H., , C. Zhang, , and J. D. Glick, 1999: Interannual variation of the Madden–Julian oscillation during austral summer. J. Climate, 12 , 25382550.

    • Search Google Scholar
    • Export Citation
  • Keen, R. A., 1982: The role of cross-equatorial tropical cyclone pairs in the Southern Oscillation. Mon. Wea. Rev., 110 , 14051416.

  • Kessler, W. S., 2001: EOF representations of the Madden–Julian oscillation and its connection with ENSO. J. Climate, 14 , 30553061.

  • Kiladis, G. N., , G. A. Meehl, , and K. M. Weickmann, 1994: Large-scale circulation associated with westerly wind bursts and deep convection over the western equatorial Pacific. J. Geophys. Res., 99 , 1852718544.

    • Search Google Scholar
    • Export Citation
  • Knutson, T. R., , and K. M. Weickmann, 1987: 30-60 day atmospheric oscillations: Composite life cycles of convection and circulation anomalies. Mon. Wea. Rev., 115 , 14071436.

    • Search Google Scholar
    • Export Citation
  • Lander, M. A., 1990: Evolution of the cloud pattern during the formation of tropical cyclone twins symmetrical with respect to the equator. Mon. Wea. Rev., 118 , 11941202.

    • Search Google Scholar
    • Export Citation
  • Lau, K-M., , L. Peng, , C. H. Sui, , and T. Nakazawa, 1989: Dynamics of super cloud clusters, westerly wind bursts, 30-60 day oscillations and ENSO: A unified view. J. Meteor. Soc. Japan, 67 , 205219.

    • Search Google Scholar
    • Export Citation
  • Lengaigne, M., , J-P. Boulanger, , C. Menkes, , S. Masson, , G. Madec, , and P. Delecluse, 2002: Ocean response to the March 1997 westerly wind event. J. Geophys. Res., 107 .8015, doi:10.1029/2001JC000841.

    • Search Google Scholar
    • Export Citation
  • Lengaigne, M., , J-P. Boulanger, , C. Menkes, , G. Madec, , P. Delecluse, , E. Guilyardi, , and J. Slingo, 2003: The March 1997 westerly event and the onset of the 1997/98 El Niño: Understanding the role of the atmospheric response. J. Climate, 16 , 33303343.

    • Search Google Scholar
    • Export Citation
  • Lengaigne, M., , E. Guilyardi, , J-P. Boulanger, , C. Menkes, , P. Delecluse, , P. Inness, , J. Cole, , and J. Slingo, 2004: Triggering of El Niño by westerly wind events in a coupled general circulation model. Climate Dyn., 23 , 601620.

    • Search Google Scholar
    • Export Citation
  • Lin, X., , and R. H. Johnson, 1996: Kinematic and thermodynamic characteristics of the flow over the western Pacific warm pool during TOGA COARE. J. Atmos. Sci., 53 , 695715.

    • Search Google Scholar
    • Export Citation
  • Love, G., 1985: Cross-equatorial influence of winter hemisphere subtropical cold surges. Mon. Wea. Rev., 113 , 14871498.

  • Luther, D. S., , D. E. Harrison, , and R. A. Knox, 1983: Zonal winds in the central equatorial Pacific and El Niño. Science, 222 , 327330.

    • Search Google Scholar
    • Export Citation
  • Madden, R. A., , and P. R. Julian, 1971: Detection of a 40–50 day oscillation in the zonal wind in the tropical Pacific. J. Atmos. Sci., 28 , 702708.

    • Search Google Scholar
    • Export Citation
  • Madden, R. A., , and P. R. Julian, 1972: Description of global-scale circulation cells in the tropics with a 40–50 day period. J. Atmos. Sci., 29 , 11091123.

    • Search Google Scholar
    • Export Citation
  • Madden, R. A., , and P. R. Julian, 1994: Observations of the 40–50-day tropical oscillation—A review. Mon. Wea. Rev., 122 , 814837.

  • Marcus, S. L., , J. O. Dickey, , and O. de Viron, 2001: Links between intraseasonal (extended MJO) and ENSO timescales: Insights via geodetic and atmospheric analysis. Geophys. Res. Lett., 28 , 34653468.

    • Search Google Scholar
    • Export Citation
  • Matsuno, T., 1966: Quasi-geostrophic motions in the equatorial area. J. Meteor. Soc. Japan, 44 , 2543.

  • McPhaden, M. J., 1999: Genesis and evolution of the 1997–98 El Niño. Science, 283 , 950954.

  • McPhaden, M. J., 2004: Evolution of the 2002/03 El Niño. Bull. Amer. Meteor. Soc., 85 , 677694.

  • McPhaden, M. J., , and B. A. Taft, 1988: Dynamics of seasonal and intraseasonal variability in the eastern equatorial Pacific. J. Phys. Oceanogr., 18 , 17131732.

    • Search Google Scholar
    • Export Citation
  • McPhaden, M. J., , H. P. Freitag, , S. P. Hayes, , B. A. Taft, , Z. Chen, , and K. Wyrtki, 1988: The response of the equatorial Pacific Ocean to a westerly wind burst in May 1986. J. Geophys. Res., 93 , 1058910603.

    • Search Google Scholar
    • Export Citation
  • McPhaden, M. J., , F. Bahr, , Y. Du Penhoat, , E. Firing, , S. P. Hayes, , P. P. Niiler, , P. L. Richardson, , and J. M. Toole, 1992: The response of the western equatorial Pacific Ocean to westerly wind bursts during November 1989 to January 1990. J. Geophys. Res., 97 , 1428914303.

    • Search Google Scholar
    • Export Citation
  • McPhaden, M. J., and Coauthors, 1998: The Tropical Ocean Global Atmosphere observing system: A decade of progress. J. Geophys. Res., 103 , 1416914240.

    • Search Google Scholar
    • Export Citation
  • Moore, A. M., , and R. Kleeman, 1999: Stochastic forcing of ENSO by the intraseasonal oscillation. J. Climate, 12 , 11991220.

  • Murakami, T., , and W. L. Sumathipala, 1989: Westerly bursts during the 1982/83 ENSO. J. Climate, 2 , 7185.

  • Nitta, T., 1989: Development of a twin cyclone and westerly bursts during the initial phase of the 1986–87 El Niño. J. Meteor. Soc. Japan, 67 , 677681.

    • Search Google Scholar
    • Export Citation
  • Nitta, T., , and T. Motoki, 1987: Abrupt enhancement of convective activity and low-level westerly burst during the onset phase of the 1986–87 El Niño. J. Meteor. Soc. Japan, 65 , 497506.

    • Search Google Scholar
    • Export Citation
  • Nitta, T., , T. Mizuno, , and K. Takahashi, 1992: Multi-scale convective systems during the initial phase of the 1986/87 El Niño. J. Meteor. Soc. Japan, 70 , 447466.

    • Search Google Scholar
    • Export Citation
  • Reynolds, R. W., , N. A. Rayner, , T. M. Smith, , D. C. Stokes, , and W. Wang, 2002: An improved in situ and satellite SST analysis for climate. J. Climate, 15 , 16091625.

    • Search Google Scholar
    • Export Citation
  • Roundy, P. E., , and G. N. Kiladis, 2006: Observed relationship between oceanic Kelvin waves and atmospheric forcing. J. Climate, 19 , 52535272.

    • Search Google Scholar
    • Export Citation
  • Saji, N. H., , B. N. Goswami, , P. N. Vinayachandran, , and T. Yamagata, 1999: A dipole mode in the tropical Indian Ocean. Science, 401 , 360363.

    • Search Google Scholar
    • Export Citation
  • Seiki, A., , and Y. N. Takayabu, 2007: Westerly wind bursts and their relationship with intraseasonal variations and ENSO. Part II: Energetics over the western and central Pacific. Mon. Wea. Rev., 135 , 33463361.

    • Search Google Scholar
    • Export Citation
  • Slingo, J. M., , D. P. Rowell, , K. R. Sperber, , and F. Nortley, 1999: On the predictability of the interannual behavior of the Madden–Julian oscillation and its relationship with El Niño. Quart. J. Roy. Meteor. Soc., 125 , 583609.

    • Search Google Scholar
    • Export Citation
  • Sperber, K. R., 2003: Propagation and the vertical structure of the Madden–Julian oscillation. Mon. Wea. Rev., 131 , 30183037.

  • Sui, C-H., , and K-M. Lau, 1992: Multiscale phenomena in the tropical atmosphere over the western Pacific. Mon. Wea. Rev., 120 , 407430.

    • Search Google Scholar
    • Export Citation
  • Takayabu, Y. N., , T. Iguchi, , M. Kachi, , A. Shibata, , and H. Kanzawa, 1999: Abrupt termination of the 1997–98 El Niño in response to a Madden–Julian oscillation. Nature, 402 , 279282.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., 1997: The definition of El Niño. Bull. Amer. Meteor. Soc., 78 , 27712777.

  • Vecchi, G. A., 2006: The termination of the 1997–98 El Niño. Part II: Mechanisms of atmospheric change. J. Climate, 19 , 26472664.

  • Vecchi, G. A., , and D. E. Harrison, 2000: Tropical Pacific sea surface temperature anomalies, El Niño, and equatorial westerly wind events. J. Climate, 13 , 18141830.

    • Search Google Scholar
    • Export Citation
  • Yu, L., , and M. M. Rienecker, 1998: Evidence of an extratropical atmospheric influence during the onset of the 1997–98 El Niño. Geophys. Res. Lett., 18 , 35373540.

    • Search Google Scholar
    • Export Citation
  • Yu, L., , R. A. Weller, , and W. T. Liu, 2003: Case analysis of a role of ENSO in regulating the generation of westerly wind bursts in the western equatorial Pacific. J. Geophys. Res., 108 , 31283147.

    • Search Google Scholar
    • Export Citation
  • Zhang, C., , and J. Gottschalck, 2002: SST anomalies of ENSO and the Madden–Julian oscillation in the equatorial Pacific. J. Climate, 15 , 24292445.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Histogram for zonal (surface–10 m) wind anomaly averaged between 2.5°N and 2.5°S (bar) and normal distribution (contour) over (a) the Indian Ocean, (b) the western and central Pacific, and (c) the eastern Pacific for the period of 1 Jan 1979–31 Aug 2002. The anomalies are defined as deviations from the 91-day running-mean climatology.

  • View in gallery

    Time–longitude section of monthly SST (°C) averaged between 5°N and 5°S in shades and WWBs in circles derived from (a) ERA-40 data (1979–Aug 2002), (b) SSM/I data (Jul 1987–2001), and (c) TOGA TAO data (1992–2002). Circles indicate the days and longitudes of maximum anomalies, with larger ones representing strong WWBs with anomalies over twice the threshold. SST data for the period of 1979–Nov 1981 are not available.

  • View in gallery

    Longitudinal distribution of accumulated WWB occurrence numbers for the entire analysis period of 1979–Aug 2002, binned to every 5° (e.g., 90°E indicates the longitudes of 90° and 92.5°E).

  • View in gallery

    Monthly distribution of accumulated WWB occurrences over (a) the Indian Ocean, (b) the western Pacific, (c) the central Pacific, and (d) the eastern Pacific. Time series are represented from July to June.

  • View in gallery

    Time–longitude section of WWBs for a 3-yr composite overlaid on the composite daily zonal wind averaged for 5°N–5°S. The composite reference is the five El Niño years (1982, 1986, 1994, 1997, and 2002) starting from July. The 3-yr time series are presented from pre–El Niño years to post–El Niño years. Larger circles indicate WWBs with over 10 m s−1 anomalies and smaller ones 5–10 m s−1. Black arrows and blue lines represent the domains and boundaries of the four regions, respectively. Two dotted–dashed lines indicate boundaries among pre–El Niño–El Niño–post–El Niño years. The period from Jul 1990 to Jun 1993 is excluded because of its unusual El Niño development.

  • View in gallery

    Lag correlation between 5-month running-mean WWB frequencies and 5-month running-mean SST anomalies in the Niño-3 region. WWB frequencies are counted over (a) the Indian Ocean, (b) the western Pacific, (c) the central Pacific, and (d) the eastern Pacific. Negative lag represents WWBs preceding SSTAs. Significance levels are shown by solid lines (99%), dotted–dashed lines (95%), and dotted–dotted–dashed lines (90%).

  • View in gallery

    Map of average locations of composite references based on WWBs. Cross marks indicate average longitudes for the reference points, which are 79.7°E, 145.6°E, 167.6°E, and 160.5°W over the Indian Ocean, the western, the central, and the eastern Pacific, respectively.

  • View in gallery

    Composite maps of SST on day 0 for WWBs over (a) the Indian Ocean, (b) the western Pacific, (c) the central Pacific, and (d) the eastern Pacific. SST is represented in contours with intervals of 1.0°C and dotted–dashed contours represent 29.5°C. Shaded regions indicate more than the 95% significance level. The abscissa is relative longitude (RLO) and 0° RLO represents where each WWB attains its maximum zonal wind anomaly.

  • View in gallery

    Same as in Fig. 8 but for composite of 91-day running-mean (RM91d) surface wind fields. The running-mean zonal winds are shown in shades with the 95% significance level with red contours. Vectors represent the composite RM91d wind field where either the zonal or meridional component is significant at the 95% level.

  • View in gallery

    Same as in Fig. 9 but for composites of the 20–100-day bandpass-filtered OLR and surface wind anomaly field. Composite OLR is shown in shades, and contours indicate the 95% significance level. Composite wind anomaly fields relative to the RM91d climatology are shown in vectors where either the zonal or meridional component is significant at the 95% level. Numbers on the right side show lags in days from the reference day (i.e., day 0) when WWBs attain their maximum amplitudes.

  • View in gallery

    Time–longitude section of composite bandpass-filtered OLR averaged between 5°N and 5°S for WWBs over (a) the Indian Ocean, (b) the western Pacific, (c) the central Pacific, and (d) the eastern Pacific. Composite OLR is shown in contours with intervals of 2.0 W m−2, where solid lines represent negative values and dotted–dashed lines show positive ones. Shaded regions indicate more than the 95% significance level. Numbers on the left side show lags in days from the reference day (i.e., day 0) when WWBs attain the maximum amplitudes.

  • View in gallery

    Same as in Fig. 11 but for a composite of bandpass-filtered zonal winds at 200 hPa averaged between 5°N and 5°S. Composite U is shown in contours with an interval of 0.5 m s−1, where solid lines represent negative values and dotted–dashed lines show positive ones. Shaded regions indicate more than the 95% significance level.

  • View in gallery

    Time series of 3-month running-mean WWB frequencies (thick lines) with labels on the left ordinate and 3-month running mean MJO indices (thin lines) on the right ordinate. The MJO index is defined as the variance (106 m2 s−2) of the bandpass-filtered (20–100 days) velocity potential at 200 hPa averaged between 10°N and 10°S over each region. The WWB frequencies and the MJO index are counted and averaged, respectively, over (a) the Indian Ocean, (b) the western Pacific, (c) the central Pacific, and (d) the eastern Pacific.

  • View in gallery

    Histograms for MJO amplitudes normalized with their std devs over (a) the Indian Ocean, (b) the western Pacific, (c) the central Pacific, and (d) the eastern Pacific, binned to every 0.5 std devs. Bar graphs indicate the number of MJO events with WWBs (black) and without WWBs (gray). Line graph indicates the ratio of the events with WWBs in each bin.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 242 242 51
PDF Downloads 206 206 73

Westerly Wind Bursts and Their Relationship with Intraseasonal Variations and ENSO. Part I: Statistics

View More View Less
  • 1 Institute of Observational Research for Global Change, Japan Agency for Marine–Earth Science and Technology, Yokosuka, Kanagawa, Japan
  • | 2 Center for Climate System Research, University of Tokyo, Kashiwa, Chiba, and Institute of Observational Research for Global Change, Japan Agency for Marine–Earth Science and Technology, Yokosuka, Kanagawa, Japan
© Get Permissions
Full access

Abstract

Statistical features of the relationship among westerly wind bursts (WWBs), the El Niño–Southern Oscillation (ENSO), and intraseasonal variations (ISVs) were examined using 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis data (ERA-40) for the period of January 1979–August 2002. WWBs were detected over the Indian Ocean and the Pacific Ocean, but not over the Atlantic Ocean. WWB frequencies for each region were lag correlated with a sea surface temperature anomaly over the Niño-3 region. WWBs tended to occur in sequence, from the western to eastern Pacific, leading the El Niño peak by 9 months to 1 month, respectively, and after around 11 months, over the Indian Ocean. These results suggest that WWB occurrences are not random, but interactive with ENSO. Composite analysis revealed that most WWBs were associated with slowdowns of eastward-propagating convective regions like the Madden–Julian oscillation (MJO), with the intensified Rossby wave response. However, seasonal and interannual variations in MJO amplitude were not correlated with WWB frequency, while a strong MJO event tended to bear WWBs. It is suggested that the strong MJO amplitude promotes favorable conditions, but it is not the only factor influencing WWB frequency. An environment common to WWB generation in all regions was the existence of background westerlies around the WWB center near the equator. It is inferred that ENSO prepares a favorable environment for the structural transformation of an MJO, that is, the intensified Rossby wave response, that results in WWB generations. The role of the background wind fields on WWB generations will be discussed in a companion paper from the perspective of energetics.

Corresponding author address: Ayako Seiki, Institute of Observational Research for Global Change, JAMSTEC, 2-15 Natsushimachou, Yokosuka, Kanagawa 237-0061, Japan. Email: aseiki@jamstec.go.jp

Abstract

Statistical features of the relationship among westerly wind bursts (WWBs), the El Niño–Southern Oscillation (ENSO), and intraseasonal variations (ISVs) were examined using 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis data (ERA-40) for the period of January 1979–August 2002. WWBs were detected over the Indian Ocean and the Pacific Ocean, but not over the Atlantic Ocean. WWB frequencies for each region were lag correlated with a sea surface temperature anomaly over the Niño-3 region. WWBs tended to occur in sequence, from the western to eastern Pacific, leading the El Niño peak by 9 months to 1 month, respectively, and after around 11 months, over the Indian Ocean. These results suggest that WWB occurrences are not random, but interactive with ENSO. Composite analysis revealed that most WWBs were associated with slowdowns of eastward-propagating convective regions like the Madden–Julian oscillation (MJO), with the intensified Rossby wave response. However, seasonal and interannual variations in MJO amplitude were not correlated with WWB frequency, while a strong MJO event tended to bear WWBs. It is suggested that the strong MJO amplitude promotes favorable conditions, but it is not the only factor influencing WWB frequency. An environment common to WWB generation in all regions was the existence of background westerlies around the WWB center near the equator. It is inferred that ENSO prepares a favorable environment for the structural transformation of an MJO, that is, the intensified Rossby wave response, that results in WWB generations. The role of the background wind fields on WWB generations will be discussed in a companion paper from the perspective of energetics.

Corresponding author address: Ayako Seiki, Institute of Observational Research for Global Change, JAMSTEC, 2-15 Natsushimachou, Yokosuka, Kanagawa 237-0061, Japan. Email: aseiki@jamstec.go.jp

1. Introduction

In recent decades, numerous observational and modeling studies have been performed to elucidate El Niño–Southern Oscillation (ENSO) variability because of its serious effects on the global climate, and a broad picture has begun to emerge. It has been clarified that ENSO is a coupled atmosphere–ocean oscillation, dominated primarily by the interaction between large-scale equatorial oceanic wave processes and variations in atmospheric circulation (e.g., Bjerknes 1969). However, it remains difficult to explain the comprehensive aspects of ENSO, especially for its vastly varying temporal characteristics and amplitudes among individual ENSO events. Even if the ocean possesses preferable conditions for El Niño, such as heat content buildup along the equator, it is not sufficient to predict the El Niño development (McPhaden 2004). This is attributable to the fact that secondary impacts of high-frequency synoptic-scale atmospheric variations were ignored.

Westerly wind bursts (WWBs), synoptic-scale disturbances that occur near the equator, are considered to have a large impact on El Niño. Previous studies have indicated that these disturbances can exert an influence on large-scale variation, namely the development of El Niño, through directly exciting oceanic Kelvin waves that propagate eastward in the Pacific along the equator (Harrison and Schopf 1984; McPhaden et al. 1988; McPhaden et al. 1992; McPhaden 1999; Lengaigne et al. 2002). Thus, it is important to investigate how and when synoptic-scale disturbances, such as WWBs, are developed in order to forecast ENSO variability. Such effects of WWBs on changes in sea surface temperature (SST) anomalies in the Pacific have been statistically depicted (Vecchi and Harrison 2000).

Concerning the atmospheric phenomena in the generation process of WWBs and its relation to ENSO, it was suggested that WWBs, in association with twin cyclones over the western Pacific, accelerated the development of the ENSO warm phase in 1986–87 (Nitta and Motoki 1987; Nitta 1989). Keen (1982) pointed out that the distribution of cyclone pairs corresponds with the eastward shift of high SST in the Pacific, using 10 yr of data from 1970 to 1979. A case study was conducted by Lander (1990) on equatorial convection evolving into twin cyclones. Luther et al. (1983) concluded that the weakening of trade winds over the central Pacific before El Niño, from 1950 to1978, was related to a series of strong WWBs. In addition, Murakami and Sumathipala (1989) emphasized that collective occurrences of WWBs lasting 7–20 days over the western Pacific were related to ENSO.

Numerical studies using intermediate coupled models have demonstrated that an impact of stochastic WWBs upon ENSO depends on both background state and ENSO phases (Moore and Kleeman 1999; Fedorov 2002; see also a summary in Fedorov et al. 2003). On the other hand, Eisenman et al. (2005), using an intermediate coupled model and observations, reported that WWB occurrences are not stochastic, rather they are modulated by ENSO. Numerical results using an atmospheric general circulation model also suggested strong impacts of WWBs at the onset and during the development of El Niño, not only by the direct impact of WWBs on the ocean, but also by their indirect influence on atmospheric variability (Lengaigne et al. 2003). These theories were confirmed using a coupled general circulation model, though the sensitivity was large (Lengaigne et al. 2004). Thus, the relationship between WWBs, especially those over the Pacific, and ENSO has been examined in various studies. However, the global distribution of equatorial WWBs and their relationship to ENSO have not yet been clarified.

Hartten (1996) identified many occurrences of WWBs over the western Pacific. In addition, Harrison and Vecchi (1997) showed the composite results of WWB occurrences over eight regions in the western Pacific, and in some region, WWBs were negatively correlated with Southern Oscillation index (SOI). These two works stated that one of four phenomena—a single cyclone, cyclones in both hemispheres, a twin cyclone, or cross-equatorial flow—were associated with WWBs. Prior to these two studies, Harrison and Giese (1991) showed latitudinal characteristics of WWBs using historical island data and concluded that WWBs on or south of the equator were the strongest.

Previous studies described some cases of WWB generations associated with an enhancement of eastward-propagating Madden–Julian oscillation (MJO; Madden and Julian 1971, 1972, 1994), resulting in twin cyclones (Nitta et al. 1992; Sui and Lau 1992; Lin and Johnson 1996). Lau et al. (1989) reported a hierarchical structure among intraseasonal variations (ISVs), super–cloud clusters, and WWBs using idealized numerical experiments. However, a statistical relationship between WWBs and the MJO has not yet been clarified.

It has also been proposed that WWBs are caused by midlatitude forcing such as cold surges over the western Pacific (Love 1985; Chu 1988; Kiladis et al. 1994). Yu and Rienecker (1998) indicated in their case study of the 1997–98 El Niño that WWBs were associated with cyclones induced by a northerly surge in phase with the convective passage of the MJO. Two case analyses showed that changes in northerly surge pathways, influenced by ENSO phases, were related to WWB occurrences through cyclone formations over the western Pacific (Yu et al. 2003). However, it remains statistically unclear whether the surges are essential for WWB occurrences and the possibility that the MJO prepares the favorable environment for intrusion of cold surges.

As for the relationship between ENSO and MJO, it was shown that ISVs with periods of 20–100 days, which is the usual MJO frequency, are uncorrelated with the ENSO cycle (Slingo et al. 1999; Hendon et al. 1999). However, longer periods of ISVs of 60–100 days (Marcus et al. 2001), or the third EOF of ISVs combined with lower modes (Kessler 2001), were correlated with the ENSO cycle. Moreover, intraseasonal oceanic Kelvin waves, mainly excited by atmospheric wind stress, have a dominant period of approximately 70 days, which is longer than the period of atmospheric ISVs (McPhaden and Taft 1988; Hendon et al. 1998). These discrepancies are debatable topics. Roundy and Kiladis (2006) indicated a transient decrease in phase speeds of oceanic Kelvin waves coupled to atmospheric disturbances as oceanic conditions tend toward El Niño. This was suggested to be partly responsible for these discrepancies.

Zhang and Gottschalck (2002) indicated a tendency for larger SST anomalies of ENSO warm events in the eastern Pacific, to be preceded by stronger oceanic Kelvin wave anomalies induced by the MJO in the western Pacific. On the other hand, Takayabu et al. (1999) suggested that a convectively coupled atmospheric Kelvin wave accelerated the termination of the 1997–98 El Niño. These authors indicated that convectively coupled ISVs propagated along the equator in May 1998 and had a Kelvin wave structure in the lower troposphere, but a Gill-type structure in the upper troposphere. Thus, further studies about variations of MJO structures affecting El Niño development and termination are important. Meridional changes in convection in response to the seasonal movement in SST were shown to be another source of the abrupt termination (Vecchi 2006).

The aim of this paper is to statistically analyze the features of WWBs over the equatorial latitudes at all longitudes, and to look for environmental conditions favorable for WWB generations. We focus on their relationships with ENSO and MJO. In a companion paper (Seiki and Takayabu, hereafter Part II), we will examine the energetics of synoptic-scale disturbances in the generation process of WWBs under different phases of ENSO and the MJO.

The data are described in section 2. Section 3 is devoted to the definition of WWBs. In sections 4 and 5, general distributions and seasonal variations in WWBs are examined, respectively. In section 6, we perform a lag correlation analysis to examine the relationship between ENSO and WWBs. In section 7, a composite analysis is performed to investigate the environmental conditions favorable for WWB generation. The relationship between MJO amplitude and WWB frequency is examined in section 8. The final section provides discussion and a conclusion.

2. Data

Primary data used for the identification of WWBs were the mean daily values of 10-m winds, recorded four times daily, on 2.5° × 2.5° grids for the period of January 1979–August 2002 from the 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis data (ERA-40). Special Sensor Microwave Imager (SSM/I) global ocean surface wind data provided by the Physical Oceanography Distributed Active Archive Center (PODAAC) of the National Aeronautics and Space Administration (NASA), and surface wind observations from Tropical Atmosphere–Ocean (TAO) buoys by Tropical Ocean Global Atmosphere (TOGA) Program in the Pacific provided by the National Oceanic and Atmospheric Administration/Pacific Marine Environmental Laboratory (NOAA/PMEL) were used to assess the quality of the ERA-40 winds. Surface wind data from SSM/I were available for July 1987–December 2001 with a resolution of 1° × 1°. SSM/I wind observations are characterized by high resolution and high coverage, but lack the wind direction information. Therefore, their wind directions were supplemented by ECMWF analysis data (Atlas et al. 1996). Daily surface winds, observed by approximately 70 TOGA TAO buoys, cover the equatorial band between 9°N and 8°S in latitude, 137°E and 95°W in longitude (McPhaden et al. 1998). For the TOGA TAO winds, we used data from 1992 to 2002 because of their better coverage during this period. Daily outgoing longwave radiation (OLR) data derived by NOAA on 2.5° × 2.5° grids for the period of 1979–2002 were also used as a proxy for convective activity in the Tropics. Weekly 1° × 1° gridded Reynolds optimal interpolation sea surface temperature (OISST) data (Reynolds et al. 2002) from 1981 to 2002 were also used. Both 200-hPa zonal winds (ERA-40) and OLR were bandpass filtered with half-power frequency cutoffs at 20 and 100 days to examine the intraseasonal variations discussed in section 7b.

3. Definition of WWBs

First, we identify WWBs. Harrison and Vecchi (1997) used anomalous surface winds relative to the monthly climatology to identify WWBs. Hartten (1996) used a WWB criterion in which 1000-hPa zonal winds exceeded 5 m s−1 with a zonal extent over 10° and lasting 2 days. On the other hand, Murakami and Sumathipala (1989) used bandpass-filtered 850-hPa zonal winds and identified WWBs with abrupt accelerations of westerly winds. Although there are various definitions for WWBs proposed in the previous studies, we used a modified version of the Hartten (1996) definition, that is, the anomalous zonal winds from the seasonal climatology. We used this definition because we aimed at detecting WWBs at all equatorial longitudes. The climatological zonal winds near the equator have large longitudinal variations. In addition, there are seasonal variations. Therefore, we define WWB events based on the following three conditions.

First, we constructed a 91-day running-mean daily climatology using 10-m zonal wind data for 1979–2001 to obtain climatological seasonal variation, that is, periods longer than the ISV. Then, anomalies were defined as the deviations from this daily climatology. The 91-day average was based on a biannual variation predominant in the equatorial Tropics. Note that longitudinal and seasonal distributions were robust in the sensitivity test with 61-day running means. Hereafter, “wind anomaly” represents the deviation from the 91-day running mean. WWBs were defined as cases that satisfied the following three criteria.

  1. Surface zonal wind anomalies averaged between 2.5°N and 2.5°S met or exceeded 5 m s−1. This meridional width for the mean values was chosen with reference to the oceanic radius of deformation, since we implicitly consider the excitation or amplification of oceanic Kelvin waves as a primary effect of WWBs.
  2. The area that satisfied the above zonal wind criteria extended zonally over at least 10° in longitude, corresponding to approximately 1100 km along the equator.
  3. The above two conditions lasted for at least 2 days.

The last continuity criterion is met when the distance between the maximum-anomaly longitudes of two continuous days are 7.5° or closer. These criteria mainly refer to Hartten (1996), except for the use of anomalous values to exclude the effects of seasonal and longitudinal variations in zonal winds.

The standard deviation for all zonal wind anomalies averaged for 2.5°N–2.5°S was 1.72 m s−1. The proportion of anomalous winds exceeding a 5 m s−1 threshold was 1.3% for all longitudes, and 1.9% of the data for the focused domain (40°E–80°W), as will be introduced in the next section. Therefore, our first criterion picks out extreme cases. Note that the total maximum zonal wind velocities of all WWBs were all westerly winds throughout the study region at this threshold.

Figure 1 shows histograms for zonal wind anomalies and the corresponding normal distributions in three regions: the Indian Ocean (40°–100°E), the western/central Pacific (120°E–180°), and the eastern Pacific (180°–100°W). Here, the zonal wind anomalies over the western and central Pacific were counted together because of the narrow extent for the central Pacific. It is interesting to note that the wind anomalies over the Indian Ocean have a near-normal distribution (Fig. 1a). On the other hand, the curves are positively skewed, giving a longer tail for westerly wind anomalies over the Pacific. Differences between westerly and easterly wind anomalies over ±5 m s−1 in each region are noteworthy; that is, strong westerly anomalies were found more frequently than were strong easterly anomalies especially over the Pacific. Standard deviations for zonal wind anomalies were 2.05, 2.28, and 1.67 m s−1 for the Indian Ocean, the western/central Pacific, and the eastern Pacific, respectively.

4. General features of WWBs

In this section, we describe general statistics of WWB occurrences. Figure 2 shows time–longitude distributions of WWB events overlaid on SST averaged between 5°N and 5°S. To assess the adequacy of using 10-m zonal wind data from the ERA-40 reanalysis to detect WWBs, we compared the WWB distributions derived from ERA-40 data (Fig. 2a), those from SSM/I data for the period from July 1987 to 2001 (Fig. 2b), and those from TOGA TAO data from 1992 to 2002 (Fig. 2c). The standard deviations for both the SSM/I and TOGA TAO surface zonal wind anomalies were larger than those for the ERA-40 daily means by a factor of 1.2. Note that standard deviations were compared for the same period and region. The standard deviations for ERA-40 wind anomalies at 1200 UTC were larger than those for the daily mean data by a factor of only 1.05. This indicates that zonal winds in ERA-40 data were smoother than those derived from satellite or buoys. Therefore, we chose the criteria for WWBs as 5.9 m s−1 for SSM/I and TOGA TAO winds to retain the same relative anomaly with the ERA-40 winds, when normalized with standard deviations. At a glance, the spatial and temporal distributions of WWBs in ERA-40 and in the SSM/I data corresponded well (Figs. 2a and 2b). However, some differences were found, such as WWBs over the Indian Ocean in late 1987 and 1999. The WWB distributions of the TOGA TAO data were well matched with those in ERA-40 (Figs. 2a and 2c), though the TOGA TAO data were obtained only in the Pacific and with low spatial resolutions, about every 10°–15° in longitude. From the above comparisons and the availability of long-term data, we decided that the ERA-40 wind data were adequate for identifying WWBs in this study.

All WWBs were detected over the Indian Ocean or the Pacific Ocean, but not over the Atlantic Ocean. Moreover, temporal variations in interannual time scales were stronger over the Pacific than those over the Indian Ocean. During ENSO warm events, WWBs occurred frequently over the Pacific, particularly to the east of 160°E, where WWB occurrences corresponded to the highest SST regions. This is consistent with the findings of Luther et al. (1983) and Keen (1982). It is noteworthy that the distribution of WWBs extended as far as 100°W during the two largest warm events in 1982–83 and 1997–98.

A total longitudinal frequency distribution of WWBs binned to every 5° of longitude is shown in Fig. 3. Here, longitudinal 5° bins were designed such that 90°E bin included those WWBs detected at 90°E and at 92.5°E, for example. As seen in Fig. 3, occurrences of WWBs were distributed over two longitudinal regions, from 60° to 92.5°E and from 135°E to 102.5°W. Interestingly, WWBs were not observed between 95° and 132.5°E, from the Maritime Continent to the western Pacific east of New Guinea, or over the Atlantic. Based on these results, the study domain is classified into four regions with reference to the WWB longitudinal distribution: the Indian Ocean (40°–100°E), the western Pacific (120°–155°E), the central Pacific (155°E–180°), and the eastern Pacific (180°–100°W). The border between the central and eastern Pacific was determined at 180° because WWBs east of 180° were scarcely observed except for during El Niño years, which will be shown in the following section. The numbers of WWBs over the Indian and the Pacific Oceans were 51 and 193, respectively. Those over the western, central, and eastern Pacific were 34, 87, and 72, respectively. The averaged durations of WWB events in each region were 3.14, 3.50, 4.24, and 4.14 days over the Indian Ocean and the western, central, and eastern Pacific, respectively. Over the Indian Ocean, the largest number of WWBs was observed at 85°E. Over the Pacific, the primary peak was located at 170°E and the secondary at 145°E. These longitudinal distributions were robust to the choice of threshold values of the zonal winds for WWB detections, which varied between 4 and 6 m s−1.

5. Seasonal variations and their relationship to interannual variations

In this section, we examine the seasonal features of WWBs and their relationship to ENSO. It is often pointed out that El Niño is phase locked to calendar years with its peak usually at the end of the calendar year. Therefore, to present seasonal variations in WWBs and their relationship to El Niño effectively, a “year” in this study was chosen to be from July to the following June. Figure 4 shows the accumulated monthly number of WWB events for each region.

Over the Indian Ocean (Fig. 4a), semiannual increases in WWB occurrences were found with peaks in November and from May to July, whereas WWBs seldom occurred in January or August. Over the western Pacific (Fig. 4b), WWBs were frequently found in December and from April to May. WWBs over the central Pacific (Fig. 4c) occurred the most frequently in December, and increased gradually from September onward. In addition, there were few WWBs in June and July, and WWB peaks occurred in March and August. Over the eastern Pacific (Fig. 4d), there was a predominant season from November to December. Frequent WWBs in the boreal winter, found both in the central and eastern Pacific, were consistent with the previous results of Keen (1982) and Harrison and Vecchi (1997).

To examine the relationship between WWB occurrences and seasonal and interannual variations associated with El Niño, we next classified the study period into four classes of years beginning in July: pre–El Niño (pre-EN) years, El Niño (EN) years, post–El Niño (post-EN) years, and other years. This classification basically follows the definition of El Niño periods given by the Japan Meteorological Agency (JMA), based on a 5-month running-mean SST anomaly averaged between 4°N–4°S and 150°–90°W (which is nearly the same as the Niño-3 region). El Niño years defined by JMA in our analysis period, were 1982–83, 1986–87, 1991–92, 1992–93, 1997–98, and from July to August 2002. We excluded an unusual intermittent El Niño in 1991–93 because it was difficult to define pre-EN and post-EN years. We added 1994–95 to our El Niño years because the 1994–95 index was similar to that during El Niño years and was classified as an El Niño year in Trenberth (1997). Pre-EN and post-EN years were determined as the years beginning in July both before and after EN years, respectively.

Figure 5 shows a 3-yr Hovmöller diagram of composite WWBs plotted over composite total zonal winds averaged between 5°N and 5°S, referring the five El Niño years.

The most apparent feature in Fig. 5 is the longitudinal expansion of WWB distribution from the western Pacific (140°E) around March in pre-EN years to the eastern Pacific (110°W) around January in EN years. From August to December in EN years, during which weak or moderate El Niño conditions were present, WWBs aggregate around the date line from 160°E to 150°W associated with enhanced westerly winds in the background over the central Pacific. It is noticed that most of the WWBs over the eastern Pacific occurred during EN years. This is consistent with the suggestion by Luther et al. (1983) that frequent WWBs over the eastern Pacific are observed during El Niño. It is notable that almost all WWBs found in the western Pacific (120°–155°E), shown in Fig. 4b from March to May, occurred during pre-EN years, while WWBs did not occur to the east of the date line except in June. In post-EN years, there were fewer WWBs over the Pacific than there were in EN or pre-EN years.

Over the Indian Ocean (40°–100°E), as illustrated in Fig. 4a, a semiannual cycle of WWB occurrences can be observed with peaks in November and around May. In Fig. 5, we can observe that WWB occurrences in these seasons are found in pre-EN and post-EN years, but not during EN years. Increases in WWBs around November and May were associated with seasonal intensifications of background westerly winds around the equator between 60° and 100°E during the monsoon transition periods. However, during EN years, background westerlies were weaker than they were in pre-EN or post-EN years for those seasons and WWBs were scarcely found.

Although the above classification of years includes some subjectivity, an interesting seasonal relationship between ENSO and WWB occurrences is found in this composite. It is known that WWBs occur most frequently between November and April (Keen 1982; Harrison and Vecchi 1997). However, this composite suggests that such seasonal change depends on interannual variation. Some case studies have indicated that WWBs in May for the 1986–87 El Niño and in March during the 1997–98 El Niño triggered an eastward migration of a warm SST (Nitta et al. 1992; McPhaden 1999). It is noteworthy to show statistically that frequent WWBs are observed from March to May over the western Pacific in pre-EN years, but not over the Pacific in post-EN years. In addition, a semiannual increase in WWB occurrences over the Indian Ocean was influenced by an El Niño phase.

6. The correlation between WWB frequencies and ENSO

In this section, we quantitatively examine the relationship between WWBs and ENSO. For ENSO indices, SST anomalies (SSTAs) averaged in the Niño-3 (5°N–5°S, 150°–90°W) region were used. We calculated correlation coefficients between 5-month running-mean WWB frequencies and 5-month running-mean Niño-3 SSTAs. Here, WWBs were counted separately in the four oceanic regions, that is, the Indian Ocean and the western, central, and eastern Pacific.

Figure 6 shows lag correlations between WWB frequencies and the ENSO indices for each region. As for WWBs over the western Pacific (Fig. 6b), there was a strong positive correlation between WWB frequency and ENSO, with a peak value of 0.62 leading Niño-3 SSTAs by 9 months. This indicates that WWBs were most likely observed over the western Pacific at approximately 9 months prior to the El Niño peak. Correlations exceeded the 99% significance level from lag −14 to −5 months. Similar features were found for the central Pacific, but the timing was different (Fig. 6c). The largest positive correlation with a value of 0.73 was found 3 months in advance of the Niño-3 SSTAs, and significant correlations beyond the 99% level were found from lag −8 to lag +2 months. Over the eastern Pacific (Fig. 6d), WWB frequencies have the strongest positive correlation coefficient of 0.78 at lag −1 month with Niño-3 SSTAs, and significant correlations exceeded the 99% level from lag −6 to +5 months.

For the Indian Ocean (Fig. 6a), WWB frequency had the largest positive correlation coefficient (0.36), lagging the Niño-3 SSTAs with 11 months, which approached the 99% significance level. A negative correlation with Niño-3 SSTAs of 0.34 was found at lag −3 months, but it was not significant at the 99% significance level. Marginally significant negative correlations, beyond the 95% significance level, existed continuously before the El Niño peak from lag −8 to 0 months. Thus, WWBs frequently occurred in phase with La Niña peaks, with a negative correlation around lag 0, or 11 months following El Niño peaks over the Indian Ocean. These results are consistent with the results in Fig. 5; that is, WWBs seldom occurred over the Indian Ocean during El Niño years. Here, further analyses are required to tell the relationships between the 11 months following El Niño and the peak of La Niña. At least, it is apparent that WWBs over the Indian Ocean occur most frequently under La Niña conditions, not under El Niño conditions.

Previous studies have indicated that WWBs occurring over the western Pacific excite Kelvin waves in the ocean and encourage the onset of El Niño (Harrison and Schopf 1984; McPhaden et al. 1988; McPhaden et al. 1992; McPhaden 1999; Lengaigne et al. 2002). Referring to these studies, our results suggest that WWB occurrences over the western and central Pacific trigger or accelerate El Niño, and WWBs over the eastern Pacific maintain El Niño conditions. WWBs over the Indian Ocean occur during La Niña though their impact on ENSO is uncertain.

On the other hand, the different timings of WWBs among all regions suggest that ENSO prepares a favorable environment for WWB generation, and the sequence of WWBs from the western to the eastern Pacific infers the eastward migration of a favorable environment controlled by ENSO. In contrast to some previous assumptions of stochastic WWBs (Moore and Kleeman 1999; Fedorov 2002), our results suggest that the interrelationship between WWBs and the ENSO life cycle is not a stochastic, but an interactive one. This is consistent with the inferences from recent studies (Lengaigne et al. 2003, 2004; Eisenman et al. 2005; Roundy and Kiladis 2006).

7. Composite results based on WWBs

In this section, we examine the environmental conditions of WWB occurrences, and their relationship to the MJO.

We made composites of SST, surface winds, and 20–100-day bandpass-filtered (hereafter referred to as bpf-) OLR referring to WWB occurrences. The points at which each WWB event attained its maximum amplitude on the longitude–time section of equatorial zonal wind anomaly data were chosen as composite reference points. Composites were made for relative days from −30 to +30 and for relative longitudes, hereafter referred to as RLO, from −180° to +180° from the references. For example, 0° RLO on day 0 represents the longitude and the day, at which equatorial WWBs attain the maximum zonal wind anomalies. Therefore, reference points were chosen over longitudinal ranges of 60°–90°E, 135°–152.5°E, 155°–177.5°E, and 180°–105°W over the Indian Ocean, and the western, central, and eastern Pacific, respectively (see Fig. 3). Figure 7 illustrates the average locations of composite references base points (0° RLO) for each region. They are 79.7°E, 145.6°E, 167.6°E, and 160.5°W over the Indian Ocean, and the western, central, and eastern Pacific, respectively. Statistical significances were examined using a Student’s t test, and we analyzed the composite values satisfying a 95% confidence level.

a. Basic state

We begin by looking at large-scale environmental conditions related to the generation of WWBs, such as SST and wind fields with 91-day running means (RM91d wind fields).

Figures 8 and 9 show the composite SST and RM91d surface winds, respectively, over each region. We show maps only for day 0 composites because time evolution was seldom found for these fields. Because of the continental distribution, the amount of SST data available for this composite varies among composite grid points. Therefore, in this analysis, composite results of SST were shown when data were available for more than 60% of total reference points.

Over the western and central Pacific (Figs. 8b and 8c), both composite centers were located in the regions above 29°C SST. The highest (>29.5°C) SST regions were found east of the composite center, from +10° to +40° RLO in the western Pacific, and from 0° to +20° RLO in the central Pacific. It is notable that environmental SST distributions in the two Pacific regions were similar, even though the actual longitudes for the composite center differed by 22°. Thus, WWBs over the western and the central Pacific occurred over the western part of the highest SST area. Under these conditions, background westerly winds penetrated into the easterly trade wind region near the composite center from −40° to +10° RLO over the equator, for both regions (Figs. 9b and 9c).

Over the eastern Pacific (Figs. 8d and 9d), a high SST area exceeding 29°C was from −50° to +15° RLO. Although the highest SST region was found to the west of the composite center, an SST area above 29°C was found around the composite center as seen in the western and central Pacific. Background westerlies lie from −40° to −10° RLO. These statistical results for SST in the Pacific are consistent with the previous case study of 1997–98 El Niño, which indicated that WWBs were apparent only over waters SSTs warmer than about 29°C (McPhaden 1999). It is notable that environmental conditions among the three Pacific regions resembled each other more than those inferred from the average reference-point longitudes. This result supports the idea that ENSO phases influence the common environmental conditions associated with WWB generation as mentioned in section 6.

In contrast, over the Indian Ocean (Fig. 8a), the warm SST region was found farther to the east of the composite center than it was over the Pacific. The area of above 29°C SST extended from the eastern Indian Ocean to the western Pacific, in average. The highest SST region was located around +60° RLO, which approximately corresponds to the north of New Guinea in the western Pacific. While SST distributions differed from those in the Pacific, the background wind fields (Fig. 9a) were more similar with respect to the westerlies. Strong background westerlies were found from −15° to +10° RLO north of the equator around the composite center. These results suggest that the enhancement of equatorial background westerlies during monsoon transition periods is of primary importance for the WWB occurrences over the Indian Ocean. These similarities about background wind fields in all four oceanic regions are noteworthy. In Part II, the effects of these background westerlies on WWB generation processes will be pursued.

b. Relationship to the intraseasonal variation

Next, we examined the relationship between WWB occurrences and the MJO. To represent the MJO, 20–100-day bpf-OLR and bpf-zonal winds at 200 hPa were used. The composite method is the same as that described in the previous section, but here we consider time series of composite results.

Figure 10 shows composite maps of bpf-OLR with wind anomaly vectors for WWBs occurring over each region. With respect to WWBs over the Indian Ocean (Fig. 10a), convectively suppressed regions, in pink shades, were found first around the composite center, roughly between −20° and +60° RLO on day −20. Then, convective activities, in blue shades, became larger to the west of the composite center, simultaneously with the weakening and eastward shift of the suppressed region on day −15. Convective activity near the composite center was amplified from day −10, in associatation with the appearance of northeasterly winds in the Northern Hemisphere around −15° RLO, and with a wide-spreading enhancement of easterly wind anomalies to the east of composite center from +40° to +120° RLO around the equator. A twin-cyclone pattern in both convective activities and a wind anomaly field appeared around the composite center on day 0 when WWBs attained the maximum amplitudes. This wind pattern resembled the structure of the so-called Matsuno–Gill pattern in response to the equatorial heat source (Matsuno 1966; Gill 1980), though the convective activity was tightly coupled to Rossby wave modes, and the Kelvin wave is already leaving the heat source. After day 0, convective regions moved away from the equator to the Northern and Southern Hemispheres with a weakening of westerly winds on the equator. In turn, a suppressed region appeared again on day +10 around the composite center.

This alternate appearance of convectively active and suppressed regions is shown clearly in Fig. 11a, which depicts a time–longitude section of composite bpf-OLR averaged between 5°N and 5°S. A convective region propagated eastward from days −20 to −10, and then intensified as it slowed down around the composite center. The most active convection period was observed at approximately 3 days prior to the maximum amplitude of the WWBs. Then, the convective region began to propagate eastward again after day 0 with a reduced amplitude. Before and after the convective region, suppressed regions propagated in a similar manner.

Over the western Pacific (Fig. 10b), an eastward propagation of a convective region, which was clearer than that over the Indian Ocean, was found from day −20 to day −5. There was an indication of northeasterly winds in the Northern Hemisphere around −30° RLO on day −15 and these northeasterlies shifted eastward around −10° RLO on day −10, accompanying the eastward migration of the convective region. These northeasterly winds may infer an influence of cold surge from the midlatitudes on WWB generation over the western Pacific, as emphasized in previous studies (Love 1985; Chu 1988; Kiladis et al. 1994; Yu and Rienecker 1998). However, the time lags between the surges and WWBs were greater than a few days, which was also mentioned in previous studies. Further studies are needed to reveal the necessity of surges for WWB occurrences and the effect of MJO on these surges. Then, from day −10 to day 0, the convective region became tighter and stronger, simultaneous to a slowdown in the eastward propagation speed, and to an amplification of the suppressed region to the west. Easterly wind anomalies to the east of the composite center were observed until day 0. However, they were weaker than those in the Indian Ocean and did not show an eastward propagation. This difference seems consistent with the MJO characteristics indicated in previous studies (Hendon and Salby 1994; Sperber 2003), as a weaker subsidence to the east of the convective region was observed over the western Pacific compared to the Indian Ocean. On day 0, a closed cyclone north of the equator and a cyclonic flow south of the equator were observed. A convective region was separated into the Northern and Southern Hemispheres on day 5, but the one in the Northern Hemisphere remained more active. This asymmetry across the equator might have been affected by the presence of Papua New Guinea, located south of the equator, considering the average longitude of 145.6°E for the composite center. This island is believed to prevent the enhancement of cyclonic circulation south of the equator, but further studies are needed. Another notable feature is the existence of a strongly suppressed region that was located to the west following the convective region from day −10 to day +10 (Fig. 10b), actually to day +25 (Fig. 11b).

Similar features were also found in the central Pacific (Fig. 10c). However, the easterly anomalies to the east of the strong convective region, found in the previous two regions, were not distinct here. Stationary westerly wind anomalies were found around the composite center from day −20 corresponding to positive SST associated with the El Niño phase shown in Fig. 8c. Convection became activated to the south of the equator on day −10 accompanying southerly winds from −20° to 0° RLO south of the equator. Then, twin cyclones appeared around the composite center on day 0, and the eastward propagation of the convective region ceased. It appeared that as the convective disturbances approached the stationary westerly wind anomaly region, the convection was enhanced, resulting in the WWB generation. Noteworthy is that convective regions did not propagate farther eastward after day 0 over the western and central Pacific, which can be clearly seen in Figs. 11b and 11c.

For each of three regions, both the Rossby wave response and the Kelvin wave response to the convective forcing were found over the Indian Ocean, whereas the Kelvin wave response was absent after day 0 over the western and central Pacific. It is suggested that this absence of the Kelvin wave response disrupts the farther eastward propagation of the convective region over the western and central Pacific.

Figures 10d and 11d are the composites for the WWBs over the eastern Pacific. Westerly winds turned from northeasterlies and southeasterlies in the Northern and Southern Hemisphere, respectively, and blew into a convective region near the equator around −30° RLO from day −20 to day −15. Similar to those over the western and central Pacific, an eastward-propagating convective region and the following convective suppressed region were found. However, the significant convective region was narrower, and a twin cyclone around the composite center on day 0 was not clearly observed (Fig. 10d). The convective peak coincided with the maximum amplitude of the westerly wind anomalies on day 0 as shown in Fig. 11d, whereas it preceded the peak of the westerly wind anomalies by a few days over the other three regions.

These results strongly indicate that the generation of WWBs is associated with the eastward propagation of the MJO-like convection over all regions. The relationship between the WWBs and the eastward-propagating MJO was more evident in Hovmöller diagrams of composite bpf-zonal winds at 200 hPa averaged between 5°N and 5°S (Fig. 12). Contrary to bpf-OLR, bpf-upper zonal winds continued propagating eastward even over the Pacific after day 0 (Figs. 12b–d). This indicates that upper disturbances continue to propagate eastward after convective activities decay over the Pacific. The maximum values of the bpf-upper easterly winds precede those of the bpf-OLR by a quarter phase, consistent with the characteristics of MJOs described by Knutson and Weickmann (1987), except for that over the eastern Pacific. In addition, an apparent periodicity of about 40 days, found both in convectively suppressed and active regions (Fig. 11) and in westerly and easterly wind regions at upper levels (Fig. 12), also corresponded to the MJO characteristics. Slowdowns of propagation were not found clearly in the upper zonal wind fields. Therefore, it is considered that WWBs originate from the MJO with its transformation of the structure at low levels, and finally separate from the MJO itself.

8. The relationship between MJO amplitude and WWB frequencies

So far in this study, significant lag correlations were found between the WWB frequency and ENSO. On the other hand, WWB generation was associated with MJO-like eastward-propagating convective disturbances. Thus, there remains a question why interannual variations in MJO amplitude are not correlated with ENSO (Slingo et al. 1999; Hendon et al. 1999). In this section, we examine the relationship between WWB frequency and MJO amplitude.

First, the relationship between the MJO amplitude and WWB frequencies in seasonal and interannual variations was examined. To represent the seasonal and interannual behaviors of the MJO, 3-month running means were applied to the WWB frequencies, and to the MJO index, defined with the total variance of bandpass-filtered (20–100 days) χ at 200 hPa averaged between 10°N and 10°S in each region.

Figure 13 shows time series of the mean WWB frequencies and the mean MJO indices. Notable is that there was no correlation between the WWB frequency and the MJO index in the seasonal and interannual variations. Correlation coefficients for these two variables were 0.0045, 0.11, 0.15, and 0.032, which are all less than the 90% significance level, over the Indian Ocean and the western, central, and eastern Pacific, respectively. In contrast, there was a significant relationship between the WWB frequencies and ENSO as shown in section 6. In addition, WWBs seldom occurred during El Niño over the Indian Ocean. However, interannual variations in MJO indexes were more random. As a consequence, we hypothesize that the MJO amplitude itself was not directly related to seasonal and interannual variations in WWB frequencies.

The previous section shows that each WWB occurrence was associated with MJO convection. Next, we investigated the relationship between individual MJO amplitudes and WWBs. Figure 14 shows histograms for the normalized amplitudes of MJO events with and without WWBs binned to every 0.5 standard deviations (hereafter, referred to as std dev) for each region. The histograms were plotted in the following manner. The amplitudes of MJO events were expressed with the MJO index defined above. Then, the amplitudes were normalized with their std dev in each region. Although the ranges of average longitudes were different among the four regions, it is still meaningful to use normalized bpf-χ200, which represents the large-scale structure with wavenumbers 1–2, for investigating individual amplitudes of the MJO convection. The events with amplitudes less than 0.25 std dev were not counted as MJO events because their minima were unclear. Although, technically, weak events may not be referred to as usual “MJOs,” we call them MJOs for convenience in this analysis. Then, we defined “MJO events with WWBs” if WWB occurrences were found preceding or following an MJO event within 10 or 20 days, respectively in order to extract WWB events directly linked to the MJO. Lag days were determined subjectively, but based on the MJO time scale and the aforementioned composite results (Fig. 11), indicating that peaks of active MJO convection were found preceding WWBs by a few days. Other events were classified into “MJO events without WWBs.” More than 90% of the total WWB events were associated with MJOs according to this definition.

The largest number of events without WWBs was found at 0.5–1.0 std devs over all regions, and their number decreased with an increase in overall amplitude. In contrast, the largest number of events with WWBs was found at 1.0–1.5 std devs among all regions. The ratio of events with WWBs increased with amplitudes up to at least 2.5 std dev. Although the ratio above 2.5 std dev does not increase monotonously because of fewer samples, a relatively higher ratio was found with stronger events. Another notable point is that the ratio of events with WWBs scarcely exceeded 50%. These results indicate that strong MJO events tend to bear WWBs but not always, suggesting that the strong MJO amplitude is a favorable condition, but not a sufficient condition for WWB generation. It is considered that the lack of significant correlations between the MJO amplitude and the WWB frequency was attributed to the existence of another controlling factor, that is, the large-scale environmental field prepared by ENSO, which acts on strong MJO events to bear WWBs.

9. Summary and discussion

In this study, we first defined WWBs over the equatorial latitudes at all longitudes derived from zonal wind anomalies using 10-m wind data from ERA-40 for the period of January 1979–August 2002. Then, observational features of WWBs and their relationship with MJO and ENSO were examined. WWBs were found either over the Indian Ocean (51) and the Pacific Ocean (193), with none over the Atlantic Ocean.

Significant lag correlations existed between the WWB frequency and ENSO at different times in each region. Keen (1982) and Luther et al. (1983) indicated that an abrupt increase and an eastward migration of WWBs were found over the Pacific during El Niño phase. A relevant study can be found in Harrison and Vecchi (1997) showing that the SOI was lag correlated with the preceding WWB occurrences over the western Pacific by 1 yr, and correlated simultaneously with those over the central Pacific, but not over the eastern Pacific. Zhang and Gottschalck (2002) indicated that Kelvin wave forcing, associated with MJO over the western Pacific, preceded El Niño by 6–12 months.

In our study, sequential occurrences of WWBs across the entire Pacific, from the west to the east, demonstrated the global variation of the WWB frequency affected by the ENSO phase. In addition, significant occurrences, which are out of phase with El Niño, were observed over the Indian Ocean. These results do not support the theory that WWBs are randomly occurring phenomena that stochastically accelerate the El Niño development as suggested by Moore and Kleeman (1999) and Fedorov (2002). Instead, our results statistically support the idea that WWB occurrences are greatly influenced by large-scale environmental variations associated with ENSO (Lengaigne et al. 2003, 2004; Eisenman et al. 2005; Roundy and Kiladis 2006).

To identify background factors that govern the development of WWBs, we performed a composite analysis based on the longitude of WWB occurrences. In the WWB generation, background wind fields were similar in all areas, even though SST distributions differed between the Indian and the Pacific Oceans. The warm pool was located near WWBs over the Pacific as mentioned in McPhaden (2004), whereas it was to the east of the WWB center over the Indian Ocean. On the other hand, background total westerly maxima were generally around a composite center near the equator over all regions. It is inferred that this similar distribution of background wind field plays a crucial role for WWB occurrences over the warm ocean.

Over the Indian Ocean, seasonal distributions of WWBs are associated with variations in background westerlies. The high frequency of WWBs, around May and November, corresponds to monsoon transition periods when equatorial westerlies are intensified. The intensity of background westerlies is influenced by ENSO as well as the monsoon, and no WWB was observed even in November in the El Niño years when the background westerlies are weaker than they were in normal years. The influence of an Indian dipole mode (Saji et al. 1999) on background winds is left for future studies.

Under these equatorial background westerlies, WWB occurrences were associated with MJO over all regions. Convective regions of MJO, which propagate eastward, slowed down with an intensification when WWBs attained their maximum amplitudes. At that time, the Matsuno–Gill pattern with the Kelvin wave response to the east and the intensified Rossby wave response to the west (Matsuno 1966; Gill 1980) appeared over the Indian Ocean, whereas only the Rossby wave response appeared over the three Pacific regions. After the WWB peak, a convective region began to propagate eastward again over the Indian Ocean, whereas the eastward propagation ceased over the Pacific. The slowdown of MJO with an intensification of the Rossby wave response is a common feature associated with WWB occurrences over all regions. This result is consistent with a transient decrease in the phase speeds of oceanic Kelvin waves coupled to atmospheric disturbances with El Niño development (Roundy and Kiladis 2006). Thus, it has been suggested that WWBs are generated in association with the transformation of MJO convection, such as the intensified Rossby wave response, over the background westerly region near the equator. Considering these conditions for WWB generation, the reason for suppressed WWBs over the Atlantic might be found in the convectively decoupled characteristics of the MJO in the Western Hemisphere (Knutson and Weickmann 1987).

Another notable point is that there is no correlation between variations in WWB frequency and MJO amplitude with respect to seasonal and interannual time scales. However, individual MJO events with strong amplitude have a tendency to bear a WWB, but not always. We suggest that the strong MJO amplitude is a favorable condition, but not a sufficient condition for WWB generation. Referring also to the fact that MJOs occur irrespective of ENSO phase (Slingo et al. 1999), it is strongly inferred that ENSO controls the favorable environment in which WWBs break out through transformations of the strong MJO. The distribution of background wind fields is suggested to be a crucial factor. The mechanism for the interactions between synoptic WWBs events and interannual ENSO variations through the role of the background westerlies will be addressed in Part II of this work.

Finally, as for the relationship between MJO and ENSO, it is interesting to suggest that deceleration of the MJO propagation speed found in WWB occurrences can explain the paradox; that is, the ISV with longer periods (60–100 days) relative to the usual MJOs (Marcus et al. 2001), or the third EOF of the ISV combined with lower modes (Kessler 2001), is correlated with the ENSO cycle, and the dominant oceanic Kelvin waves period is centered at 70 days (McPhaden and Taft 1988; Hendon et al. 1998), whereas the ISV with the usual MJO frequency (20–100 days) is uncorrelated with the ENSO cycle (Slingo et al. 1999; Hendon et al. 1999). Air–sea coupling of the longer-period ISV and the ocean during El Niño development indicated by Roundy and Kiladis (2006) is an interesting topic left for future studies.

Acknowledgments

This work was part of the first author’s doctoral dissertation. The lead author would like to express thanks to both Prof. Masahide Kimoto and Prof. Akimasa Sumi, Center for Climate System Research, University of Tokyo, for their helpful comments and discussions. She is also thankful to Prof. Jun Matsumoto, University of Tokyo, and Prof. Hiroshi Niino, Ocean Research Institute, University of Tokyo, for their valuable comments to improve the manuscript. She is thankful as well to Dr. Toru Nozawa, National Institute for Environmental Studies, and Dr. Masayoshi Ishii, Meteorological Research Institute, for their help in obtaining and processing the ERA-40 data. The authors would also like to acknowledge Prof. Chidong Zhang and Dr. Paul E. Roundy for their discussions on this work.

REFERENCES

  • Atlas, R., , R. N. Hoffman, , S. C. Bloom, , J. C. Jusem, , and J. Ardizzone, 1996: A multiyear global surface wind velocity dataset using SSM/I wind observations. Bull. Amer. Meteor. Soc., 77 , 869882.

    • Search Google Scholar
    • Export Citation
  • Bjerknes, J., 1969: Atmospheric teleconnections from the equatorial Pacific. Mon. Wea. Rev., 97 , 163172.

  • Chu, P-S., 1988: Extratropical forcing and the burst of equatorial westerlies in the western Pacific: A synoptic study. J. Meteor. Soc. Japan, 66 , 549564.

    • Search Google Scholar
    • Export Citation
  • Eisenman, I., , L. Yu, , and E. Tziperman, 2005: Westerly wind bursts: ENSO’s tail rather than the dog? J. Climate, 18 , 52245238.

  • Fedorov, A. V., 2002: The response of the coupled tropical ocean–atmosphere to westerly wind bursts. Quart. J. Roy. Meteor. Soc., 128 , 123.

    • Search Google Scholar
    • Export Citation
  • Fedorov, A. V., , S. L. Harper, , S. G. Philander, , B. Winter, , and A. Wittenberg, 2003: How predictable is El Niño? Bull. Amer. Meteor. Soc., 84 , 911920.

    • Search Google Scholar
    • Export Citation
  • Gill, A. E., 1980: Some simple solutions for heat-induced tropical circulation. Quart. J. Roy. Meteor. Soc., 106 , 447462.

  • Harrison, D. E., , and P. S. Schopf, 1984: Kelvin wave induced anomalous advection and the onset of surface warming in El Niño events. Mon. Wea. Rev., 112 , 923933.

    • Search Google Scholar
    • Export Citation
  • Harrison, D. E., , and B. S. Giese, 1991: Episodes of surface westerly winds as observed from islands in the western tropical Pacific. J. Geophys. Res., 96 , 32213237.

    • Search Google Scholar
    • Export Citation
  • Harrison, D. E., , and G. A. Vecchi, 1997: Westerly wind events in the tropical Pacific. J. Climate, 10 , 31313156.

  • Hartten, L. M., 1996: Synoptic settings of westerly wind bursts. J. Geophys. Res., 101 , 1699717019.

  • Hendon, H. H., , and M. L. Salby, 1994: The life cycle of the Madden–Julian oscillation. J. Atmos. Sci., 51 , 22252237.

  • Hendon, H. H., , B. Liebmann, , and J. D. Glick, 1998: Oceanic Kelvin waves and the Madden–Julian oscillation. J. Atmos. Sci., 55 , 88101.

    • Search Google Scholar
    • Export Citation
  • Hendon, H. H., , C. Zhang, , and J. D. Glick, 1999: Interannual variation of the Madden–Julian oscillation during austral summer. J. Climate, 12 , 25382550.

    • Search Google Scholar
    • Export Citation
  • Keen, R. A., 1982: The role of cross-equatorial tropical cyclone pairs in the Southern Oscillation. Mon. Wea. Rev., 110 , 14051416.

  • Kessler, W. S., 2001: EOF representations of the Madden–Julian oscillation and its connection with ENSO. J. Climate, 14 , 30553061.

  • Kiladis, G. N., , G. A. Meehl, , and K. M. Weickmann, 1994: Large-scale circulation associated with westerly wind bursts and deep convection over the western equatorial Pacific. J. Geophys. Res., 99 , 1852718544.

    • Search Google Scholar
    • Export Citation
  • Knutson, T. R., , and K. M. Weickmann, 1987: 30-60 day atmospheric oscillations: Composite life cycles of convection and circulation anomalies. Mon. Wea. Rev., 115 , 14071436.

    • Search Google Scholar
    • Export Citation
  • Lander, M. A., 1990: Evolution of the cloud pattern during the formation of tropical cyclone twins symmetrical with respect to the equator. Mon. Wea. Rev., 118 , 11941202.

    • Search Google Scholar
    • Export Citation
  • Lau, K-M., , L. Peng, , C. H. Sui, , and T. Nakazawa, 1989: Dynamics of super cloud clusters, westerly wind bursts, 30-60 day oscillations and ENSO: A unified view. J. Meteor. Soc. Japan, 67 , 205219.

    • Search Google Scholar
    • Export Citation
  • Lengaigne, M., , J-P. Boulanger, , C. Menkes, , S. Masson, , G. Madec, , and P. Delecluse, 2002: Ocean response to the March 1997 westerly wind event. J. Geophys. Res., 107 .8015, doi:10.1029/2001JC000841.

    • Search Google Scholar
    • Export Citation
  • Lengaigne, M., , J-P. Boulanger, , C. Menkes, , G. Madec, , P. Delecluse, , E. Guilyardi, , and J. Slingo, 2003: The March 1997 westerly event and the onset of the 1997/98 El Niño: Understanding the role of the atmospheric response. J. Climate, 16 , 33303343.

    • Search Google Scholar
    • Export Citation
  • Lengaigne, M., , E. Guilyardi, , J-P. Boulanger, , C. Menkes, , P. Delecluse, , P. Inness, , J. Cole, , and J. Slingo, 2004: Triggering of El Niño by westerly wind events in a coupled general circulation model. Climate Dyn., 23 , 601620.

    • Search Google Scholar
    • Export Citation
  • Lin, X., , and R. H. Johnson, 1996: Kinematic and thermodynamic characteristics of the flow over the western Pacific warm pool during TOGA COARE. J. Atmos. Sci., 53 , 695715.

    • Search Google Scholar
    • Export Citation
  • Love, G., 1985: Cross-equatorial influence of winter hemisphere subtropical cold surges. Mon. Wea. Rev., 113 , 14871498.

  • Luther, D. S., , D. E. Harrison, , and R. A. Knox, 1983: Zonal winds in the central equatorial Pacific and El Niño. Science, 222 , 327330.

    • Search Google Scholar
    • Export Citation
  • Madden, R. A., , and P. R. Julian, 1971: Detection of a 40–50 day oscillation in the zonal wind in the tropical Pacific. J. Atmos. Sci., 28 , 702708.

    • Search Google Scholar
    • Export Citation
  • Madden, R. A., , and P. R. Julian, 1972: Description of global-scale circulation cells in the tropics with a 40–50 day period. J. Atmos. Sci., 29 , 11091123.

    • Search Google Scholar
    • Export Citation
  • Madden, R. A., , and P. R. Julian, 1994: Observations of the 40–50-day tropical oscillation—A review. Mon. Wea. Rev., 122 , 814837.

  • Marcus, S. L., , J. O. Dickey, , and O. de Viron, 2001: Links between intraseasonal (extended MJO) and ENSO timescales: Insights via geodetic and atmospheric analysis. Geophys. Res. Lett., 28 , 34653468.

    • Search Google Scholar
    • Export Citation
  • Matsuno, T., 1966: Quasi-geostrophic motions in the equatorial area. J. Meteor. Soc. Japan, 44 , 2543.

  • McPhaden, M. J., 1999: Genesis and evolution of the 1997–98 El Niño. Science, 283 , 950954.

  • McPhaden, M. J., 2004: Evolution of the 2002/03 El Niño. Bull. Amer. Meteor. Soc., 85 , 677694.

  • McPhaden, M. J., , and B. A. Taft, 1988: Dynamics of seasonal and intraseasonal variability in the eastern equatorial Pacific. J. Phys. Oceanogr., 18 , 17131732.

    • Search Google Scholar
    • Export Citation
  • McPhaden, M. J., , H. P. Freitag, , S. P. Hayes, , B. A. Taft, , Z. Chen, , and K. Wyrtki, 1988: The response of the equatorial Pacific Ocean to a westerly wind burst in May 1986. J. Geophys. Res., 93 , 1058910603.

    • Search Google Scholar
    • Export Citation
  • McPhaden, M. J., , F. Bahr, , Y. Du Penhoat, , E. Firing, , S. P. Hayes, , P. P. Niiler, , P. L. Richardson, , and J. M. Toole, 1992: The response of the western equatorial Pacific Ocean to westerly wind bursts during November 1989 to January 1990. J. Geophys. Res., 97 , 1428914303.

    • Search Google Scholar
    • Export Citation
  • McPhaden, M. J., and Coauthors, 1998: The Tropical Ocean Global Atmosphere observing system: A decade of progress. J. Geophys. Res., 103 , 1416914240.

    • Search Google Scholar
    • Export Citation
  • Moore, A. M., , and R. Kleeman, 1999: Stochastic forcing of ENSO by the intraseasonal oscillation. J. Climate, 12 , 11991220.

  • Murakami, T., , and W. L. Sumathipala, 1989: Westerly bursts during the 1982/83 ENSO. J. Climate, 2 , 7185.

  • Nitta, T., 1989: Development of a twin cyclone and westerly bursts during the initial phase of the 1986–87 El Niño. J. Meteor. Soc. Japan, 67 , 677681.

    • Search Google Scholar
    • Export Citation
  • Nitta, T., , and T. Motoki, 1987: Abrupt enhancement of convective activity and low-level westerly burst during the onset phase of the 1986–87 El Niño. J. Meteor. Soc. Japan, 65 , 497506.

    • Search Google Scholar
    • Export Citation
  • Nitta, T., , T. Mizuno, , and K. Takahashi, 1992: Multi-scale convective systems during the initial phase of the 1986/87 El Niño. J. Meteor. Soc. Japan, 70 , 447466.

    • Search Google Scholar
    • Export Citation
  • Reynolds, R. W., , N. A. Rayner, , T. M. Smith, , D. C. Stokes, , and W. Wang, 2002: An improved in situ and satellite SST analysis for climate. J. Climate, 15 , 16091625.

    • Search Google Scholar
    • Export Citation
  • Roundy, P. E., , and G. N. Kiladis, 2006: Observed relationship between oceanic Kelvin waves and atmospheric forcing. J. Climate, 19 , 52535272.

    • Search Google Scholar
    • Export Citation
  • Saji, N. H., , B. N. Goswami, , P. N. Vinayachandran, , and T. Yamagata, 1999: A dipole mode in the tropical Indian Ocean. Science, 401 , 360363.

    • Search Google Scholar
    • Export Citation
  • Seiki, A., , and Y. N. Takayabu, 2007: Westerly wind bursts and their relationship with intraseasonal variations and ENSO. Part II: Energetics over the western and central Pacific. Mon. Wea. Rev., 135 , 33463361.

    • Search Google Scholar
    • Export Citation
  • Slingo, J. M., , D. P. Rowell, , K. R. Sperber, , and F. Nortley, 1999: On the predictability of the interannual behavior of the Madden–Julian oscillation and its relationship with El Niño. Quart. J. Roy. Meteor. Soc., 125 , 583609.

    • Search Google Scholar
    • Export Citation
  • Sperber, K. R., 2003: Propagation and the vertical structure of the Madden–Julian oscillation. Mon. Wea. Rev., 131 , 30183037.

  • Sui, C-H., , and K-M. Lau, 1992: Multiscale phenomena in the tropical atmosphere over the western Pacific. Mon. Wea. Rev., 120 , 407430.

    • Search Google Scholar
    • Export Citation
  • Takayabu, Y. N., , T. Iguchi, , M. Kachi, , A. Shibata, , and H. Kanzawa, 1999: Abrupt termination of the 1997–98 El Niño in response to a Madden–Julian oscillation. Nature, 402 , 279282.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., 1997: The definition of El Niño. Bull. Amer. Meteor. Soc., 78 , 27712777.

  • Vecchi, G. A., 2006: The termination of the 1997–98 El Niño. Part II: Mechanisms of atmospheric change. J. Climate, 19 , 26472664.

  • Vecchi, G. A., , and D. E. Harrison, 2000: Tropical Pacific sea surface temperature anomalies, El Niño, and equatorial westerly wind events. J. Climate, 13 , 18141830.

    • Search Google Scholar
    • Export Citation
  • Yu, L., , and M. M. Rienecker, 1998: Evidence of an extratropical atmospheric influence during the onset of the 1997–98 El Niño. Geophys. Res. Lett., 18 , 35373540.

    • Search Google Scholar
    • Export Citation
  • Yu, L., , R. A. Weller, , and W. T. Liu, 2003: Case analysis of a role of ENSO in regulating the generation of westerly wind bursts in the western equatorial Pacific. J. Geophys. Res., 108 , 31283147.

    • Search Google Scholar
    • Export Citation
  • Zhang, C., , and J. Gottschalck, 2002: SST anomalies of ENSO and the Madden–Julian oscillation in the equatorial Pacific. J. Climate, 15 , 24292445.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

Histogram for zonal (surface–10 m) wind anomaly averaged between 2.5°N and 2.5°S (bar) and normal distribution (contour) over (a) the Indian Ocean, (b) the western and central Pacific, and (c) the eastern Pacific for the period of 1 Jan 1979–31 Aug 2002. The anomalies are defined as deviations from the 91-day running-mean climatology.

Citation: Monthly Weather Review 135, 10; 10.1175/MWR3477.1

Fig. 2.
Fig. 2.

Time–longitude section of monthly SST (°C) averaged between 5°N and 5°S in shades and WWBs in circles derived from (a) ERA-40 data (1979–Aug 2002), (b) SSM/I data (Jul 1987–2001), and (c) TOGA TAO data (1992–2002). Circles indicate the days and longitudes of maximum anomalies, with larger ones representing strong WWBs with anomalies over twice the threshold. SST data for the period of 1979–Nov 1981 are not available.

Citation: Monthly Weather Review 135, 10; 10.1175/MWR3477.1

Fig. 3.
Fig. 3.

Longitudinal distribution of accumulated WWB occurrence numbers for the entire analysis period of 1979–Aug 2002, binned to every 5° (e.g., 90°E indicates the longitudes of 90° and 92.5°E).

Citation: Monthly Weather Review 135, 10; 10.1175/MWR3477.1

Fig. 4.
Fig. 4.

Monthly distribution of accumulated WWB occurrences over (a) the Indian Ocean, (b) the western Pacific, (c) the central Pacific, and (d) the eastern Pacific. Time series are represented from July to June.

Citation: Monthly Weather Review 135, 10; 10.1175/MWR3477.1

Fig. 5.
Fig. 5.

Time–longitude section of WWBs for a 3-yr composite overlaid on the composite daily zonal wind averaged for 5°N–5°S. The composite reference is the five El Niño years (1982, 1986, 1994, 1997, and 2002) starting from July. The 3-yr time series are presented from pre–El Niño years to post–El Niño years. Larger circles indicate WWBs with over 10 m s−1 anomalies and smaller ones 5–10 m s−1. Black arrows and blue lines represent the domains and boundaries of the four regions, respectively. Two dotted–dashed lines indicate boundaries among pre–El Niño–El Niño–post–El Niño years. The period from Jul 1990 to Jun 1993 is excluded because of its unusual El Niño development.

Citation: Monthly Weather Review 135, 10; 10.1175/MWR3477.1

Fig. 6.
Fig. 6.

Lag correlation between 5-month running-mean WWB frequencies and 5-month running-mean SST anomalies in the Niño-3 region. WWB frequencies are counted over (a) the Indian Ocean, (b) the western Pacific, (c) the central Pacific, and (d) the eastern Pacific. Negative lag represents WWBs preceding SSTAs. Significance levels are shown by solid lines (99%), dotted–dashed lines (95%), and dotted–dotted–dashed lines (90%).

Citation: Monthly Weather Review 135, 10; 10.1175/MWR3477.1

Fig. 7.
Fig. 7.

Map of average locations of composite references based on WWBs. Cross marks indicate average longitudes for the reference points, which are 79.7°E, 145.6°E, 167.6°E, and 160.5°W over the Indian Ocean, the western, the central, and the eastern Pacific, respectively.

Citation: Monthly Weather Review 135, 10; 10.1175/MWR3477.1

Fig. 8.
Fig. 8.

Composite maps of SST on day 0 for WWBs over (a) the Indian Ocean, (b) the western Pacific, (c) the central Pacific, and (d) the eastern Pacific. SST is represented in contours with intervals of 1.0°C and dotted–dashed contours represent 29.5°C. Shaded regions indicate more than the 95% significance level. The abscissa is relative longitude (RLO) and 0° RLO represents where each WWB attains its maximum zonal wind anomaly.

Citation: Monthly Weather Review 135, 10; 10.1175/MWR3477.1

Fig. 9.
Fig. 9.

Same as in Fig. 8 but for composite of 91-day running-mean (RM91d) surface wind fields. The running-mean zonal winds are shown in shades with the 95% significance level with red contours. Vectors represent the composite RM91d wind field where either the zonal or meridional component is significant at the 95% level.

Citation: Monthly Weather Review 135, 10; 10.1175/MWR3477.1

Fig. 10.
Fig. 10.

Same as in Fig. 9 but for composites of the 20–100-day bandpass-filtered OLR and surface wind anomaly field. Composite OLR is shown in shades, and contours indicate the 95% significance level. Composite wind anomaly fields relative to the RM91d climatology are shown in vectors where either the zonal or meridional component is significant at the 95% level. Numbers on the right side show lags in days from the reference day (i.e., day 0) when WWBs attain their maximum amplitudes.

Citation: Monthly Weather Review 135, 10; 10.1175/MWR3477.1

Fig. 11.
Fig. 11.

Time–longitude section of composite bandpass-filtered OLR averaged between 5°N and 5°S for WWBs over (a) the Indian Ocean, (b) the western Pacific, (c) the central Pacific, and (d) the eastern Pacific. Composite OLR is shown in contours with intervals of 2.0 W m−2, where solid lines represent negative values and dotted–dashed lines show positive ones. Shaded regions indicate more than the 95% significance level. Numbers on the left side show lags in days from the reference day (i.e., day 0) when WWBs attain the maximum amplitudes.

Citation: Monthly Weather Review 135, 10; 10.1175/MWR3477.1

Fig. 12.
Fig. 12.

Same as in Fig. 11 but for a composite of bandpass-filtered zonal winds at 200 hPa averaged between 5°N and 5°S. Composite U is shown in contours with an interval of 0.5 m s−1, where solid lines represent negative values and dotted–dashed lines show positive ones. Shaded regions indicate more than the 95% significance level.

Citation: Monthly Weather Review 135, 10; 10.1175/MWR3477.1

Fig. 13.
Fig. 13.

Time series of 3-month running-mean WWB frequencies (thick lines) with labels on the left ordinate and 3-month running mean MJO indices (thin lines) on the right ordinate. The MJO index is defined as the variance (106 m2 s−2) of the bandpass-filtered (20–100 days) velocity potential at 200 hPa averaged between 10°N and 10°S over each region. The WWB frequencies and the MJO index are counted and averaged, respectively, over (a) the Indian Ocean, (b) the western Pacific, (c) the central Pacific, and (d) the eastern Pacific.

Citation: Monthly Weather Review 135, 10; 10.1175/MWR3477.1

Fig. 14.
Fig. 14.

Histograms for MJO amplitudes normalized with their std devs over (a) the Indian Ocean, (b) the western Pacific, (c) the central Pacific, and (d) the eastern Pacific, binned to every 0.5 std devs. Bar graphs indicate the number of MJO events with WWBs (black) and without WWBs (gray). Line graph indicates the ratio of the events with WWBs in each bin.

Citation: Monthly Weather Review 135, 10; 10.1175/MWR3477.1

Save