El Nino, La Nina, and the Southern Oscillation

S. George Philander, James R. Holton,Renata Dmowska

• Publisher: Academic Press
• Year: 1989
• ISBN:
  9780080570983
• Topic: Sciences physiques
• Subtopic: Météorologie et climatologie

El Nino and the Southern Oscillation is by far the most striking phenomenon caused by the interplay of ocean and atmosphere. It can be explained neither in strictly oceanographic nor strictly meteorological terms. This volume provides a brief history of the subject, summarizes the oceanographic and meteorological observations and theories, and discusses the recent advances in computer modeling studies of the phenomenon.

• Includes a comprehensive and up-to-date research survey
• Discusses in detail sophisticated computer models
• Provides a clear exposition of the major problems which prevent more accurate predictions of El Nino

Chapter 1

The Southern Oscillation: Variability of the Tropical Atmosphere

1.1 Introduction

Towards the end of the nineteenth century Hildebrandsson (1897) noticed that the atmospheric pressure fluctuations at Sydney, Australia are out of phase with those at Buenos Aires, Argentina. A few years later Lockyer and Lockyer (1902a), father and son, confirmed this and estimated the period of the oscillation to be approximately 3.8 years. Their analyses of additional data from 95 stations around the world revealed that the oscillation was almost global in extent (Lockyer and Lockyer, 1902b, 1904). The map of the pressure fluctuations that appears in their paper of 1904 is, in its gross aspects, very similar to that in Fig. 1.1, which shows that the oscillation has two centers of action, one over the western tropical Pacific and eastern Indian Ocean and the other over the southeastern tropical Pacific. Fluctuations at these two centers, which are thousands of kilometers apart, are remarkably coherent and are out of phase. This is evident in Fig. 1.2, which also shows that the interannual fluctuations are very irregular in time. Sir Gilbert Walker named these fluctuations the Southern Oscillation. In collaboration with Bliss and others he established that the Southern Oscillation involves far more than a seesaw in the surface pressure difference across the Pacific Ocean. It is associated with major changes in the rainfall patterns and wind fields of the tropical Indian and Pacific Oceans and is correlated with meteorological fluctuations in other parts of the globe (Walker, 1923, 1924, 1928; Walker and Bliss, 1930, 1932, 1937). The important relation between the Southern Oscillation and sea surface temperature variations in the tropical Pacific was not discovered until the 1960s in studies by Ichiye and Petersen (1963), Berlage (1966), and Doberitz (1968). The correlations between the various parameters establish that high surface pressure over the western and low surface pressure over the southeastern tropical Pacific coincide with heavy rainfall, unusually warm surface waters, and relaxed trade winds in the central and eastern tropical Pacific. This phase of the Southern Oscillation is known as El Niño. Although some descriptions give the impression that El Niño is a temporary departure from some “normal” condition of the tropical Pacific, this is inaccurate. Normal conditions can be defined statistically, but it is clear from Fig. 1.2 that the Pacific is usually not in a “normal” state. It is either in one phase of the Southern Oscillation, known as El Niño, or in the complementary phase for which the term La Niña is apposite. During La Niña, surface pressure is high over the eastern but low over the western tropical Pacific, while trades are intense and the sea surface temperatures and rainfall are low in the central and eastern tropical Pacific.

Figure 1.1 Correlations (×10) of annual mean sea level pressure with the pressure at Darwin. Correlations exceed 0.4 in the shaded regions and are less than −0.4 in the regions with dashed lines. [From Trenberth and Shea (1987).]

Figure 1.2 Sea level pressure fluctuations between 1937 and 1983 at Tahiti (solid line) and Darwin (dotted line) in units of standard deviations for the respective stations. (Data provided by the Climate Analysis Center, NOAA, Washington, D.C.)

The terms El Niño and La Niña cover a wide range of conditions. For example, it is evident in Fig. 1.2 that the amplitudes of different El Niño episodes vary enormously. This prompted Quinn et al. (1978) to introduce four El Niño categories—strong, moderate, weak, and very weak¹—but there are still considerable differences within each category. The strong El Niño of 1972 was more intense than that of 1976 but it was of shorter duration; it lasted for approximately 18 months whereas warm conditions lingered for several years after 1976. In 1972 and 1976 El Niño evolved in a similar manner, which is described in Section 1.4, but the exceptionally intense El Niño of 1982–1983 evolved very differently, as did the one of 1941. The term El Niño clearly covers diverse phenomena, as do La Niña and the Southern Oscillation.

There are relatively brief periods when none of these terms adequately describes conditions in the tropical Pacific. For example, the pressure fluctuations at Darwin and Tahiti are sometimes uncorrelated. It is possible for an increase in surface pressure at Darwin and a simultaneous decrease at Tahiti not to coincide with the appearance of unusually warm surface waters off Peru and with heavy rainfall in the central equatorial Pacific. It is pointless to debate whether or not El Niño occurred in such years. This problem is a consequence of the imperfect correlations between various parameters in the tropical Pacific. Table 1.1 shows that the values of the correlation coefficients, though high, are not equal to one. It follows that a definition of the Southern Oscillation in terms of the pressure difference between Darwin and Tahiti, for example, will differ from definitions in terms of sea surface temperature² or rainfall. Any one definition is of course unambiguous, but a multitude of definitions serves no purpose. It is more practical to avoid strict definitions and to accept that the terms Southern Oscillation, El Niño, and La Niña are general and qualitative. They are useful in the same way that the term winter is useful even though each winter is distinct. The features that are common to different El Niño episodes need to be identified—this is done in Section 1.4—because they provide a focus from which to explain the phenomenon. There are, however, many purposes for which this focus is too diffuse. Thus the term El Niño is no substitute for a detailed description of how different parameters varied during a certain period. A prediction that El Niño will occur is of limited practical value unless it specifies how surface pressure, rainfall, sea surface temperature, and other parameters will vary over a given period.

Table 1.1

Matrix of Contemporaneous Correlations between Various Parameters in the Tropical Pacific^a Horel and Wallace (1981).

^aCorrelation coefficients are × 100. Correlations are between time series of the sea surface temperatures (SST) index for the tropical Pacific Ocean, the sea level pressure (SLP) index, the 200-mbar index, the Pacific North American (PNA) index, and rainfall indexes on various islands. The time series are for the winter seasons of at least 28 years (1951 to 1978) unless it is indicated, in parentheses, to be shorter.

^bThe SLP index measures the Tahiti–Darwin normalized sea level pressure.

^cThe 200-mbar index measures the height of the 200-mbar surface in the tropics and is proportional to the averaged temperature of the tropical troposphere.

^dThe PNA index is drawn from the teleconnection pattern in Fig. 1.26, which is based on the 700-mbar height at the four points indicated.

The Southern Oscillation at first appears complex because the large number of correlations between various parameters in different parts of the globe presents a bewildering picture. A physically plausible framework for a discussion of the correlations emerged from the seminal papers by Bjerknes (1966, 1969). The principal result is that large-scale atmospheric motion in the tropics, on time scales of weeks and longer, corresponds to direct thermal circulations. In such circulations, moisture-laden air converges onto the warmest regions of the earth’s surface where the air rises and condenses, causing those regions to have widespread cloudiness and heavy precipitation. Elsewhere subsidence of dry air from the upper troposphere forms a lid on the planetary boundary layer and prevents small cumulus clouds from growing to a size that can produce substantial rainfall. The monsoons that bring heavy rainfall to the Indian subcontinent during the summer when that region is warmer than the surrounding ocean are an example of a direct thermally driven circulation. Other examples include the meridional Hadley Circulation, in which air rises near the equator and sinks in higher latitudes, and the zonal Walker Circulation, in which air rises over the warm western tropical Pacific and sinks over the cold eastern tropical Pacific. The Southern Oscillation is a perturbation to these direct thermal circulations and is associated with fluctuations in the intensity and the positions of the regions of rising, moist air. The factors that influence the interannual movements of the convective zones—variations in sea surface temperature patterns and variations in the heating of the continents—also influence the seasonal movements of the convective zones. The Southern Oscillation and the seasonal cycle therefore have much in common.

1.2 The Seasonal Cycle

Satellite photographs such as that in Fig. 1.3 clearly show that certain regions in the tropics are bright and highly reflective. These cloudy regions, where moist air rises and condenses, are convective zones and occur principally over equatorial Africa, Central and South America, and the “Maritime Continent” of the western Pacific and southeastern Asia. Important extensions over the oceans include the east-west Intertropical Convergence Zone (ITCZ) north of the equator and the South Pacific Convergence Zone, which slopes southeastward from the western equatorial Pacific and which has a weak counter...