| 英文摘要: | The El Niño/Southern Oscillation (ENSO) is the dominant climate phenomenon affecting extreme weather conditions worldwide. Its response to greenhouse warming has challenged scientists for decades, despite model agreement on projected changes in mean state. Recent studies have provided new insights into the elusive links between changes in ENSO and in the mean state of the Pacific climate. The projected slow-down in Walker circulation is expected to weaken equatorial Pacific Ocean currents, boosting the occurrences of eastward-propagating warm surface anomalies that characterize observed extreme El Niño events. Accelerated equatorial Pacific warming, particularly in the east, is expected to induce extreme rainfall in the eastern equatorial Pacific and extreme equatorward swings of the Pacific convergence zones, both of which are features of extreme El Niño. The frequency of extreme La Niña is also expected to increase in response to more extreme El Niños, an accelerated maritime continent warming and surface-intensified ocean warming. ENSO-related catastrophic weather events are thus likely to occur more frequently with unabated greenhouse-gas emissions. But model biases and recent observed strengthening of the Walker circulation highlight the need for further testing as new models, observations and insights become available.
The impacts of anthropogenic climate change may be felt through changes in modes of natural climatic variability. ENSO is the most important year-to-year fluctuation of the climate system on the planet1, varying between anomalously cold (La Niña) and warm (El Niño) conditions. Underpinning occurrences of ENSO events is the positive feedback between trade wind intensity and zonal contrasts in sea surface temperature (SST), referred to as the Bjerknes feedback. The trade winds normally pile up warm surface water in the western Pacific while upwelling colder subsurface water in the east along the equator and off the west coast of South America. The resulting east–west surface temperature contrast reinforces an east–west air pressure difference across the basin that in turn drives the trade winds. During La Niña, the system strengthens, but during El Niño, the trade winds weaken as atmospheric pressure rises in the western Pacific and falls in the eastern Pacific. The Bjerknes feedback now operates in reverse, with weakened trade winds and SST warming tendencies along the Equator reinforcing one another. It is still not clear what sets this quasi-oscillatory behaviour, that is, whether ENSO is self-sustaining or triggered by stochastic forcing2. What is clear is that ocean and atmosphere preconditions are required3, as supported by the fundamental characteristics of the mean tropical climate such as thermal gradients and associated circulations that balance radiative heating4. These swings in temperature are accompanied by changes in the structure of the subsurface ocean, the position of atmospheric convection and associated global teleconnection patterns, severely disrupting global weather patterns5, 6, 7, 8, 9, 10, and affecting ecosystems11 and agriculture12 worldwide. During the 1982/1983 and 1997/1998 extreme El Niño events6, 8, surface warming anomalies propagated eastward in an uncharacteristic fashion13, 14, and massive surface warm anomalies in the eastern equatorial Pacific exceeding 3 °C caused an equatorward shift of the Intertropical Convergence Zone (ITCZ). Catastrophic floods occurred in the eastern equatorial region of Ecuador and northern Peru6, 8. The South Pacific Convergence Zone (SPCZ), the largest rain band in the Southern Hemisphere, shifted equatorward by up to 1,000 km (an event referred to as zonal SPCZ10), spurring floods and droughts in south Pacific countries and shifting extreme cyclones to regions normally not affected by such events10. Other impacts included floods in the southwest United States, disappearance of marine life, and decimation of the native bird population in the Galapagos Islands15. The development of the 1997/1998 extreme El Niño event was accompanied by an extreme positive Indian Ocean Dipole in boreal autumn, affecting millions of people across countries in the Indian Ocean rim. An extreme La Niña ensued in 1998/1999, generating droughts in the southwest United States and eastern equatorial Pacific regions, floods in the western Pacific and central American countries, and increased land-falling west Pacific tropical cyclones and Atlantic hurricanes7, 9, 12. In light of these massive impacts, how ENSO will respond to greenhouse warming is one of the most important issues in climate change science. The issue has challenged scientists for decades, but there has been no consensus on how ENSO amplitude and frequency may change16, 17, 18. Past studies have proceeded without specifically looking into the response of ENSO extremes, and have focused on simple metrics such as temperature variability in the eastern equatorial Pacific and linear dynamics, assuming that the characteristics of El Niño and La Niña are symmetric. Through the Coupled Model Intercomparison Phase 5 (CMIP5) process19, substantial improvement in modelling ENSO has been made18, 20, 21. There is recognition that the two opposing extremes are not mirror opposites13, 22, 23, 24, 25, 26, 27; that is, the impacts of and processes responsible for extreme El Niño and La Niña events are not symmetric14, 28, 29, 30, 31, 32. Further, the dynamics of extreme ENSO events are different from moderate events14, 31, 32, 33, and therefore the two must be examined separately in terms of their response to greenhouse warming. With this recognition, significant progress has been made in understanding the characteristics of extreme ENSO events in models and observations, as part of the observed diversity of events, such as central Pacific ENSO33, 34, 35 or ENSO Modoki36, their likely future behaviour under greenhouse conditions, and potential changes in their teleconnections. This study provides a review of these advances. We show that the frequency of ENSO extremes is expected to increase, ENSO teleconnections are likely to shift eastward, and these changes can, to a large extent, be interpreted as consequences of mean state changes.
The dynamics and properties of ENSO are closely linked to the slowly evolving background climate state of the equatorial Pacific Ocean (for example by rectifying into the mean state37, 38), which in turn affects ENSO feedback processes1, 16. The tropical Pacific is projected to change under greenhouse warming. The projection (Box 1) includes a weakening of the Walker circulation39, 40, 41, a faster warming rate in the equatorial than off-equatorial Pacific16, 39, 41, in the eastern equatorial Pacific41 and the Maritime continent than in the central Pacific, and over the ocean surface than subsurface32, 42. The warming pattern gives rise to an increase in rainfall in the equatorial Pacific, particularly in the eastern part of the basin43.
Box 1: Mean state changes and consequences.
Features associated with a weakening Walker circulation (A). The trade winds and equatorial currents weaken, the eastern equatorial Pacific warms faster than the surrounding regions, and the thermocline shallows (present-day, black curve; future, red curve). The weakening equatorial zonal currents are conducive to an increased frequency of eastward-propagating El Niño events. The faster warming in the eastern equatorial Pacific is favourable for an increased frequency of extreme El Niño events by promoting atmospheric convection. The increased occurrences of extreme El Niño are in turn conducive to an increased frequency of La Niña owing to a discharged thermocline that promotes an influence of the subsurface cool water in the central Pacific. Increasing vertical temperature gradients (B) and enhanced warming (C) over the maritime continent. These changes are additional factors that help to produce an increased frequency of extreme La Niña events, through nonlinear zonal advection and Ekman pumping.
The schematic figure shows greenhouse-induced future changes at the surface (shown only for the north Pacific) and upper-ocean along-Equator and meridional cross-sections. Greenhouse-induced changes (red arrows) to the mean Walker circulation (dashed black arrows) and mean ocean currents (cyan arrows) are indicated. Main features of changes are indicated by letters A, B, C and D.3 El Niño and La Niña events are not symmetric in spatial pattern22, 23, 24, 59 or temporal evolution13, 60, 61. Extreme El Niño features disproportionately warm maximum SST anomalies in the eastern equatorial Pacific, but the anomaly centres of weak El Niño and extreme La Niña events are situated in the central equatorial Pacific25. The anomaly centre of weak La Niña is located further towards the eastern equatorial Pacific than extreme La Niña25, 26, 31, 32. This spatial asymmetry is characterized by positive SST skewness in the eastern equatorial Pacific, but negative skewness in the central equatorial Pacific62. In addition, an extreme La Niña tends to follow an extreme El Niño27, 32, but not the other way around. A La Niña can last for more than a year, whereas El Niño events tend to terminate abruptly in late boreal winter or spring13, 60, 61. The asymmetries require at least two ENSO indices to distinguish extreme El Niño from extreme La Niña, or extreme El Niño from weak El Niño25, 26, 31, 32. The two indices may be obtained by empirical orthogonal function (EOF) analysis of SST anomalies, which deconvolves the spatiotemporal SST variability into orthogonal modes, each described by a principal spatial pattern and the corresponding principal component (PC) time series. An event may be described by an appropriately weighted superposition of the two modes. One EOF depicts strong variability in the Niño3.4 or Niño3 region25 (Fig. 1a) and the other resembles the central Pacific El Niño pattern34, 35, 36 (Fig. 1b). An extreme El Niño (red stars, Fig. 1c,f) is described by the difference between EOF1 and EOF2, or an E-index defined as (PC1 – PC2)/√2 (ref. 25), corresponding to extreme positive SST |