Development of the 1997-98 El Niño

Fig. 1. Time-longitude sections (an average over 2°S - 2°N) of anomalous (a) SST from the Reynolds analyses (weekly average), (b) SSH from TOPEX/Poseidon altimeter (10-day average), and (c) the depth of 20°C isotherm (5-day average) derived from TOGA/TAO moorings. The anomalies in (a) are departures from the 1982-1992 base period weekly means. The anomalies in (b) are deviations from a monthly-mean surface composed from T/P data spanning from 1993-1996. The anomalies in (c) are relative to monthly climatologies linearly interpolated onto 5-day intervals.

Because of its major importance to economies of many countries, ENSO has generated intense scientific interest and endeavor in the past a few decades. Tremendous progress has been made in understanding the intrinsic ENSO mechanisms and developing skills in its prediction (Busalacchi and O'Brien, 1980; Rasmusson and Carpenter, 1982; Philander et al., 1984; Zebiak and Cane, 1987). It is now known that the ENSO cycle involves active interactions between the atmosphere and the ocean, in which wind changes cause oceanic changes and changes in tropical sea surface temperature (SST) affect atmospheric circulation. Existing theories (Schopf and Suarez, 1988; Battisti, 1988) based on linear equatorial wave dynamics and tropical air-sea interaction seem to have explained well the evolving El Niño condition. However, the mechanism(s) governs the onset of El Niño is not fully understood. Forecasts of El Niño at a lead time longer than one year still face great difficulty partly due to the lack of adequate theoretical guidance. The current El Niño occurred at the time that the most advanced technology has been put into work to support the best ever Earth monitoring network. The development of this event has been recorded not only by the Tropical- Ocean-Global-Atmosphere (TOGA) Tropical-Atmosphere-Ocean (TAO) array which measures essential oceanographic and surface meteorological variables of the tropical Pacific Ocean but also by multiple satellite sensors which provides continuous, global survey of meteorological and sea surface oceanic conditions. Analyzing these datasets should prove valuable in enhancing our understanding of the cause of El Niño and in improving the model predictability. In this report, we present analysis of meteorological and oceanic observations before and during the onset phase of this El Niño.

The meteorological and oceanic conditions across the whole equatorial Pacific have changed dramatically before and after early 1997 as shown in Fig.1. During the year 1996, both SST and SSH were higher in the region west of the dateline and lower in the east, and the westward trade winds over the central and western Pacific were persistently stronger. This strengthened trade-wind forcing drove more surface warm waters flowing westward and piling up in the western basin, which resulted in enhanced zonal contrast of the SSH and SST in the equatorial Pacific Ocean. The basin-wide spatial SST and SSH structures during this period, typified by the deviations of the monthly-mean SST and SSH fields in December 1996, are illustrated in Fig. 2. It shows that the east-west SST difference was up to 2°C and the SSH zonal slope was about 10-15 cm greater within the equatorial Pacific. The pre-onset state, characterized by (i) strengthened trade winds, (ii) enhanced zonal tilt of sea level, and (iii) enhanced east-west SST difference, actually had persisted since early 1995 (the 1995 oceanic and atmospheric conditions, which are not presented in this report, were very similar to those in 1996 as shown in Figs. 1 and 2). This condition reversed abruptly after January 1997.

Fig. 2. Anomalous SSHs from TOPEX/Poseidon altimeter for (a) Dec 1996, (b) Mar 1997, and (c) Jun 1997. SSH anomalies are departures from the 1993-1996 base period monthly means. Also shown are anomalous SSTs from the Reynolds analyses for (d) Dec 1996, (e) Mar 1997, and (f) Jun 1997. SST anomalies are departures from the 1982-1992 base period monthly means.

An eastward extension of high SSH anomaly starting in late December 1996, marked the beginning of one of the greatest El Niño events of the century. The SSH anomaly, a short-lived signal as seen in Fig.1b, propagated rapidly along the equator and reached the South American coast in February 1997. One month later, another SSH anomaly with a greater amplitude developed near the western boundary and also subsequently propagated eastward. These eastward propagating signals are known as Kelvin waves. The monthly-mean SSH deviations in March 1997 (Fig. 2) show the spatial pattern of these waves. The arrival of the first Kelvin-wave group at the eastern boundary is indicated by the positive SSH anomalies along the equator in the region east of 130°W and along the east coast. Meanwhile, the second Kelvin-wave group, being formed in the western equatorial region, was located between 150°E and 130°W. These Kelvin waves were also observed by TOGA TAO moorings. The depth of 20°C isotherm, a commonly used index for thermocline variations in the tropical Pacific Ocean, reveals clearly the eastward traveling signals associated with these two groups of Kelvin waves (Fig. 1c).

Kelvin waves are usually generated directly by wind-stress forcing or result from the reflection of Rossby waves at the western boundary. A westerly wind burst forces downwelling Kelvin waves with a depression of the thermocline and so an increase in the depth of 20°C isotherm. Correspondingly, an easterly wind burst generates upwelling Kelvin waves (Cane and Sarachik, 1976; 1977; 1979). These waves can induce SST changes primarily through horizontal advection and through their impact on equatorial upwelling. In the eastern Pacific basin where both horizontal and vertical temperature gradients are considerably greater than those in the western warm pool area, Kelvin waves are much more effective in inducing SST changes. This explains why the wave signals are less clear in the SST plot as compared with the SSH plot (see Fig. 2). In fact, the SST field was affected by the first Kelvin-wave group only in a small area near the eastern boundary as to March 1997. With the subsequent arrival of the second group of Kelvin waves in April and May, the thermocline in the east was greatly depressed and the sea surface warming in the east basin became more extensive (Figs. 1 and 2) and the unstable atmosphere-ocean interaction began to be actively involved. It is quite apparent that the two groups of downwelling Kelvin waves played a key role for the onset of the current El Niño episode. The question is whether these Kelvin waves were generated by the reflection of the Rossby waves at the western boundary or by westerly wind bursts. To examine it, the temporal evolution of the zonal wind component averaged over equatorial region is plotted in Fig. 3. These winds were derived from a blend of wind speed observations from the Defense Mapping Satellite Program Special Sensor Microwave Imager (SSM/I), from ship and buoy observations and from the European Center for Medium Range Weather Forecasts (ECMWF) surface wind analyses (Atlas et al, 1993). Figure 3 reveals that westerly wind bursts (eastward) preceded over western and central Pacific at the time when Kelvin waves were generated, with the first wind burst occurring in the period around late December 1996 and the second in late February 1997. These sudden changes in the wind fields were also recorded by the NSCAT satellite sensor and TOGA TAO moored measurements (not shown). It is obvious that the Kelvin waves which led to the onset of the current El Niño were generated by the westerly wind bursts (Luther et al, 1983; Harrison and Geise, 1991).

Fig. 3. Time-longitude section (an average over 2°S - 2°N) of anomalous surface winds (m/sec) derived from a blend of SSM/I wind speeds, ship and buoy observations and ECMWF surface wind analyses. The anomalies are departures from the monthly means (1988-1996) linearly interpolated onto weekly intervals.

Yu and Rienecker (1998) in a recent study discussed the genesis of the westerly wind bursts. By analyzing newly available satellite-derived wind fields, they found that those wind bursts resulted from interactions between the tropical intraseasonal oscillation, namely the Madden-Julian Oscillation (MJO) (Madden and Julian, 1972), and northerly surges from East Asia/Western North Pacific into the tropics. The MJO is an intrinsic atmospheric fluctuation in the equatorial region and is caused by systematic eastward movement of low-frequency large-scale convection and circulation anomalies. The deep tropical convection and lower-level westerly wind anomaly in association with the passage of an MJO are mostly profound over equatorial regions of central Indian and western Pacific Oceans. Figure 4 shows satellite observations of outgoing longwave radiation (OLR) (an index of the atmospheric convection activity) during the onset phase. As is seen from Figs. 3 and 4, the two westerly wind bursts which triggered the current El Niño were associated with the MJO. The possible link between ENSO and the MJO has been previously proposed (Lau and Chan, 1988). However, it seems that the MJO alone is not sufficient to be a trigger for the current El Niño since the ocean was already preconditioned with positive height anomalies in the western Pacific during 1995-96 and the MJO was active during that period. It is possible (and even likely) that the phase of the interannual oscillation is important in the MJO-ENSO interaction. It can be seen that during 1995 and 1996 the westerly wind anomalies over the equatorial Indian Ocean propagated no further than 130°E, while during December 1996 and May 1997 they extended to the dateline and even further east and their amplitudes were also locally enhanced in the Pacific sector. Yu and Rienecker (1998) found that the intensification of the westerly anomalies over the western Pacific was influenced by the intrusion of the extratropical atmospheric disturbances, i.e., the cold surges of northerly wind anomalies from East Asia/Western North Pacific into the tropical Pacific with a period of 1-5 days (Lau et al, 1983; Sui et el, 1997). After the penetration of midlatitude cold air, a cyclone was subsequently induced just north of the equator and the westerly wind bursts were simultaneously generated. Meanwhile, a cyclone was also developed in the south of the equator under the influence of cross-equatorial winds. A cyclone pair (Keen, 1980; Harrison and Vecchi, 1997) was then formed. The combination of equatorward surges in both hemispheres prolonged and strengthened the westerly wind event so that strong enough downwelling Kelvin wave packets were excited to induce sufficient warming in the eastern basin to initiate El Niño development.

Fig. 4. Time-longitude section (an average over 5°S - 5°N) of anomalous outgoing longwave radiation (OLR). Anomalies are departures from the 1979-1995 base period pentad means.

The study of Yu and Rienecker (1998) indicates that extratropical atmospheric processes played an important role in the initial development of El Niño events. The midlatitude intrusions coincided with the passage of the MJO when low pressure prevailed over the western tropical Pacific. Questions remain as to the distinct role of the MJO in ENSO variability and the chaotic nature of the phasing of these interactions. A satisfactory answer to these questions may also provide a key to other related issues such as why ENSO is phase-locked to annual cycles and what causes the aperiodicity. The study further indicates that models entirely based on tropical air-sea interactions may not be sufficient for extended ENSO predictions.

Acknowledgments

We thank the following people for generously providing their datasets which made this study possible: David Adamec and Ken Casey for the gridded TOPEX/Poseidon SSH, Dick Reynolds for the SST analysis, Michael McPhaden for the TOGA TAO observations, and Bob Atlas and David Adamec for the SSM/I wind analysis. The OLR data are provided by the Climate Prediction Center of NOAA.

References

Atlas, R., R.N. Hoffman, and S.C. Bloom, 1993: Surface wind velocities over the ocean. Atlas of Satellite Observations Related to Global Change, Ed., R.J. Gurney, J.L. Foster, and C.L. Parkinson, 128-139, Cambridge University Press, New York.

Battisti, D.S., The dynamics and thermodynamics of a warming event in a coupled tropical atmosphere-ocean model. J. Atmos. Sci., 45, 2889-2919.

Busalacchi, A.J., and J.J. O'Brien, 1980: The seasonal variability of the tropical Pacific. J. Phys. Oceanogr., 10, 1929-1952.

Cane, M.A., 1983: Oceanographic events during El Niño. Science, 222, 1189-1195.

Cane, M.A., and E.S. Sarachik, 1976: Forced baroclinic ocean motions. I. The linear equatorial unbounded case. J. Mar. Res., 34, 629-655.

Cane, M.A., and E.S. Sarachik, 1977: Forced baroclinic ocean motions. II. The linear equatorial bounded case. J. Mar. Res., 35, 395-432.

Cane, M.A., and E.S. Sarachik, 1979: Forced baroclinic ocean motions. III. The linear equatorial basin case. J. Mar. Res., 37, 355-398.

Harrison, D.E., and B.S. Geise, 1991: Episodes of surface westerly winds as observed from islands in the western tropical Pacific. J. Geophys. Res., 96, 3221-3237.

Harrison, D.E., and G.A. Vecchi, 1997: Westerly wind events in the tropical Pacific, 1986-95. J. Climate, 10, 3131-3156.

Keen, R.A., 1982: The role of cross-equatorial cyclone pairs in the Southern Oscillation. Mon. Wea. Rev., 110, 1405 - 1416.

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, 1109-1123.

Lau, K.M., and P.H. Chan, 1988: Interannual and intraseasonal variations of tropical convection: a possible link between the 40-day mode and ENSO? J. Atmos. Sci., 45, 506-521.

Lau, K.M., C.P. Chang, and P.H. Chan, 1983: Short-term planetary-scale interactions over the Tropics and midlatitudes. Part II. Winter MONEX period. Mon. Wea. Rev., 111, 1372-1388.

Luther, D.S., D.E. Harrison and R.A. Knox, 1983: Zonal winds in the central equatorial Pacific and El Niño. Science, 222, 327-330.

Philander, S.G.H., T. Yamagata, and R.C. Pacanowski, 1984: Unstable air-sea interactions in the tropics. J. Atmos. Sci., 41, 604-613.

Schopf, P.S., and M.J. Suarez, 1988: Vacillations in a coupled ocean-atmosphere model. J. Atmos. Sci., 45, 549-566.

Sui, C.H., K.M. Lau and Y. Ge, 1997: Variability of tropical circulation and its interaction with extratropical disturbances. Proceedings, 1st International Conference on Reanalyses, Silver Spring, MD, USA, World Climate Research Programme, International Global Energy and Water Cycle Experiment Project Office.

Yu, L., and M. Rienecker, 1998: Onset of the 1997-98 El Niño. Submitted.

Zebiak, S.E., and M.A. Cane, 1987: A model ENSO. Mon. Wea. Rev., 115, 2262-2278.