CLIMPAG | CLIMATE INDICATORS | An El Niño Primer

El Niño-Southern Oscillation defined
El Niño is a local warming of surface waters which takes place in the entire equatorial zone of the central and eastern Pacific Ocean off the Peruvian coast and which affects the atmospheric circulation world-wide. It usually peaks around Christmas, hence the name of the phenomenon: El Niño is Spanish for Christ Child). La Niña refers to the "cold" equivalent of El Niño.

Like most atmospheric phenomena, it occurs at more or less regular intervals (pseudo-cycles) and, as such, there is nothing "abnormal" in the occurrence of El Niño. El Niño triggers so much interest for three reasons: it can be modelled, thus forecast; its influence on climate is global, and there is a time lag between the phenomenon itself and many of its most important climatic consequences. Therefore, it can be used for forecasting climate [1].

The Southern Oscillation is an East-West balancing movement of air masses between the Pacific and the Indo-Australian areas. It is associated (roughly synchronised) with typical wind patterns and El Niño, and "measured" by the Southern Oscillation Index (SOI) [2]. El Niño is the oceanic component, while the Southern Oscillation is the atmospheric one. This combination gives rise to the term ENSO. Although there is no perfect correlation between El Niño and the Southern Oscillation for minor variations, large negative values of the SOI are associated with warm events.

As illustrated in figure 1 below, El Niño occurs on average every 4 to 5 years, sometimes less (2 to 3 years), sometimes more (8 to 11 years). The phenomenon proper lasts 12 to 18 months, also a recent very unusual event lasted from mid-1990 to mid-1995 [3]. As indicated, El Niño is a permanent feature of the Pacific ocean; the wording warm event describes an anomalous, unusually warm El Niño. Similarly, there are cold events when the Sea Surface Temperatures (SST) become unusually cold.

Figure 1: Recent variations of the Southern Oscillation Index
Graph 1

Low (negative) values correspond to the warm phase. The graph was prepared based on time series taken from the Queensland Department of Primary Industries web-site (QDPI, Australia) and the Goddard Distributed Active Archive Center (DAAC, USA). The February 1998 value corresponds to the 30-day period ending on 25 February (given by QDPI). SOI is normalised, i.e. it is the dimensionless deviation from the 1951-80 average expressed in standard deviation units.

It is important to realise that the ENSO phenomenon has always happened and will continue to do so. What has changed recently is a better modelling and understanding of the physics, hence better forecasting and more focused monitoring of ENSO, as well as better world-wide forecasting of the climatic consequences.

There are essentially four sources for ENSO forecasts: the US National Oceanic and Atmospheric Administration (NOAA; see http://www.pmel.noaa.gov/toga-tao/el-nino/forecasts.html#enso), the Australian Bureau of Meteorology (BoM; see http://www.bom.gov.au/bmrc/mrlr/rzk/climfcn3.htm), the Lamont Doherty group at Columbia University (see http://rainbow.ldeo.columbia.edu) and the Centre for Ocean-Land-Atmosphere Studies (COLA/IGES funded by NOAA, NSF, and NASA; see http://grads.iges.org/nino/) A warm event had been forecast for 1997 by one of the climate forecasting Centres in January. However, the agreement between the forecasts could only be reached when the actual warming started to be observed. The models still disagree on details like the intensity, the timing and geographic extension of the SST peak. The warming has now been amply confirmed by satellite observations, moored and floating buoys, deep ocean temperature soundings and sea level analysis.

The phenomenon is very visible in figure 2 below, where SOI is compared with some records (1982, 1972), as well as the average of the 5 years of this century (1900 to 1996) which recorded the lowest SOI values. The value indicated for June is among the lowest ever. This behaviour is also very clear in the SST records and various other ENSO indices (see, for instance, under http://www.cdc.noaa.gov/ENSO/enso.current.html).

The current El Niño is exceptional in that it started very early in March-April and because it reached record values during the summer 1997. It actually culminated in June-August but had adopted a more normal pattern in the late months of 1997, seasonally increasing again in December-January and remaining stationary in February 1998. The "previous strongest" El Niño (1982) peaked in March-April but started slower than the current one. The 1972 event (also one of the top 6 for this century) collapsed after July [4]. Therefore, it is very difficult to say exactly how the situation will develop this time. It is clear by now that some extreme weather occurred, though not necessarily where it was expected. On the other hand, some predicted disasters did not occur. Of course, the situation is still developing, and no final conclusions can yet be drawn.

El Niño affects marine fisheries, particularly in the eastern Pacific Ocean where it tends to radically lower the ocean primary production. It is thought to be associated with declines of a number of fish stocks, including the largest fishery that has ever existed on earth, the Peruvian anchoveta fishery, which totally collapsed in conjunction with the 1972 El Niño.

El Niño has different impacts in different parts of the world, sometimes beneficial (for instance, it is believed to suppress hurricanes in the Atlantic), sometimes negative (drought in Southern Africa and Northern Latin America). Some of the possible consequences are illustrated in figure 3 below. Although for many of them the mechanisms are well understood, it should be noted that they are based on the empirical/statistical observation that extreme ENSO has been associated in the past with extreme climatic phenomena. In this context, extreme is taken in a statistical sense, i.e. rare, occurring with a low frequency and usually associated with unusual intensities [5].

Figure 2: A comparison of the current SOI development with historical data (The data used are the same as for figure 1)
Graph 2

The figure given for February 1998 is the average of the 25 first days of the month.

From an agricultural perspective, it is important to remember that the cropping season of the northern tropics falls between April and September, while it coincides with October-March in the southern hemisphere (this excludes all-year moist equatorial areas, the "sub-tropical" Mediterranean areas and winter crops of the temperate climates). It is thus essential to pay attention to the relative timing of the warm events and the crop calendar: the top of figure 3 is thus essentially post-ENSO, while figure 3b coincides with the build-up of high SST (pre-ENSO).

Several WWW sites have assessments of El Niño impacts at a regional and national level. As they have a lot of cross-referencing, it is sometimes difficult to point to the actual source of the data. Some of them are listed below, by continents.

Several WMO members have ENSO information, for instance Brazil, Chile, Colombia, France, UK and the USA. They can be found under http://www.wmo.ch/web-en/member.html; more are added regularly. Meteorology, climatology and geography departments of universities also frequently carry information on ENSO.

Figure 3: World-wide climatic impacts of warm ENSO events

The upper half (3a) corresponds to the northern hemisphere winter (October to March) and the lower half (3b) covers impacts during April to September. D indicates drought, R stands for unusually high rainfall (not necessarily unusually intense rainfall) and W indicates abnormally warm periods. The figure is modified from two illustrations given by Pacific Marine Environmental Laboratory WWW home page (NOAA, USA).

According to the analyses of the Queensland Department of primary industries (http://www.dpi.qld.gov.au/longpdk), the following areas may experience unusual rainfall conditions in the period February-April:

Details about the current anomalies can be taken from WWW site of NCEP (National Center for Environmental Prediction), CPC (Climate Prediction Center) of National Oceanic and Atmospheric Administration (NOAA), USA (http://nic.fb4.noaa.gov/products/analysis_monitoring/GLOB_CLIM/). The situation described is that of the end of September - early October. It is stressed that the anomalous weather shown is not necessarily a consequence of El Niño.

Generally wet in the North west, but dry in most other parts of the country, particularly the NE (Queensland)

In general, the timing and strengthening observed for this event in recent months has been well ahead of that seen in other warm episodes over recent decades (e.g. since the early 1950s). Although the 1982-83 warm event is still regarded as the strongest this century, the present event has already exceeded it by some measures, and may yet develop into a record-breaking event.

FAO does not have any mandate in the geo-physical aspects of the El Niño phenomenon. Our interests are only in the impacts on agriculture -and consequently food security - of the extreme negative/positive climatic events that can be triggered by El Niño. This involves essentially the monitoring of the ENSO situation, which is now easily done through a number of excellent WWW sites, for instance the International Research Institute for Climate prediction (http://iri.ucsd.edu), the Climate Diagnostic Centre (http://www.cdc.noaa.gov), the American Geophysical Union (http://earth.agu.org), the Institut Français de Recherche Scientifique pour le Développement en Coopération (http://www.orstom.fr), as well as the other WWW sites mentioned above.

Given the diversity of ENSO situations, and the resulting diversity of climate scenarios, the seasonal forecasts of the extreme climate that may result from El Niño (locally: at the level of countries and below) are the competence of the sole national meteorological services. Needless to say, regional/sub-continental centres can play an important role whenever the national meteorological services either lack the capacity or the information required to issue adequate warnings for the farming community. In southern Africa, it is the SADC Regional Early Warning Unit and the WMO Drought Monitoring Center which have introduced the ENSO problématique and operational warnings in the region (Matarira and Unganai, 1994)

There are obvious benefits in knowing in advance what a cropping season is going to be like, as this will allow farmers and governments alike to take preventive action mainly in the areas of crop management (see http://www.pmel.noaa.gov/toga-tao/el-nino/impacts-brazil.html for an excellent example) and marketing (as commodity prices can be better predicted).

The methodology of whether to take management decisions taking into account ENSO information must be based as much on economic data as on potential agronomic impacts. But it is essential that the seasonal forecasts be accompanied by reliable probabilities of occurrence which will allow the calculation of quantitative agricultural scenarios.

It would be too narrow an approach to develop a strategy that would deal only with "reacting against El Niño". The proper strategy is to develop response mechanisms that can be used at different levels (from Government to farmers) to react to short and medium-term weather forecasts (24 hours up to one week) and to more or less long-term climatic forecasts. Whether the extreme factor is due to El Niño or any other cause is not really relevant. (The only typical feature of El Niño impacts being better than average predictability).

The proper reaction to extreme ENSO events must be seen in a broader context, i.e. the development of a strategy covering extreme atmospheric factors in general and involving several institutional partners. Current discussions about how to react to El Niño are useful only if they lead to long-term solutions. It is suggested that the proper strategy would incorporate the following :

As El Niño develops, some of the expected consequences actually materialize, while others are not verified. It is clear that the current event is exceptional also in terms of the attention it received from the media, governments and the man in the street alike. Some countries were able to realistically manage the situation while others over-reacted or found themselves completely unprepared. Much can be learnt from the current situation in terms of adaptation and mitigation strategies.

1. Climate forecasts, also know as seasonal forecasts differ in many respects from weather forecasts especially as regards the input data, the methods used and the time horizon. Weather forecasts cover up to 10 days, while climate forecasts extend up to one year ahead. Another difference is that climate is defined as average weather, thus a climatic forecast is much less detailed than a weather forecast: only general situations are outlined (i.e. "dry conditions" , "abnormally warm", etc.) in comparison with a reference period.

2. The Southern Oscillation is not unlike the swinging movement of water in a bath tub. It is popular because it is easy to measure as the difference between sea level atmospheric pressures at Darwin (N Australia) and Tahiti (South- Central Pacific), for which there is a long historic record which has greatly facilitated research.

3. This can be considered as one long, chronic, multiple-peaked El Niño or rather as several small ones coming in rapid sequence. What is important is that it lacked the general periodicity of the earlier historical period.

4. There are some marked differences between the current El Niño and the July 1972 event: the SSTs stayed warm throughout 1972 (NINO3 was +2.64 in December, then cooled rapidly in 1973). From an SST viewpoint the collapse was not as rapid as with the SOI, and it is the SST anomalies that provide surplus energy for some climate impacts. NINO3 is currently at +3C and NINO1.2, along the coast is at +4.2C, hence the "feeling" of many experts that the event is not going to disappear rapidly. Note that NINO3 and NINO1.2 correspond to different "sectors" of the central Pacific ocean.

5. One of the consequences is a relative difficulty to achieve good statistical significance. It should further be noted that the frequency is referred to a time period which may differ. For instance, figures 1 and 2 adopt the 1951 to 1980 period as reference. Since warm events have been relatively more frequent after 1980, a more recent reference period would have ranked 1997 as less "extreme". Also, El Niño is only one of the factors that affect global atmospheric circulation and climate. As far as is known, it does not depend on other external forcing. Therefore, there cannot possibly be a "perfect" association between El Niño and climatic impacts. For instance, 1957/58 or 1977/78 had virtually no effect in Southern Africa. Finally, warm events differ in many ways, such as their "position" in the sequence of warm/cold/average SSTs, their absolute magnitude, duration, location, etc.

6. This can be explained, for instance, by the fact that ENSO results in a "climatic complex" characterized by rainfall, solar radiation (linked to cloudiness), temperatures, etc. All the factors contribute to the final production figure. In other words, ENSO variables may be more synthetic than a single climatic factor.

There is a very large volume of fundamental and applied research on El Niño. Most quoted WWW sites will provide an easy entry point, maybe starting from http://earth.agu.org/revgeophys. For general overviews, also refer to the web site of WMO (http://www.wmo.ch).

Not all references are quoted in the text above. Some have been added because of their focus on agriculture or because they present a good overview (e.g. Glantz, 1996).

Bryceson, K.P. and D.H. White, Editors, 1994. "Proceedings of a Workshop on Drought and Decision Support". Department of Primary Industries and Energies, Bureau of Resources Sciences Canberra, Australia. vii + 63 pp. By the people who run one of the most operational ENSO WWW sites.

Cane,M.A., G. Eshel and R.W. Buckland,1994. 'Forecasting Zimbabwean maize yield using eastern equatorial Pacific sea surface temperature'. "Nature", 370: 204-205.

Costa Amaral, O., 1993. 'A variabilidade pluviométrica no leste do nordeste do Brasil e o evento enos de 1992'. "Boletim de geografia teorética", 23(45-46): 61-70.

Glantz, M.H., 1996. "Currents of change: El Niño's impact on climate and society". Cambridge University Press UK, xiii + 194 pp.

Instituto de fomento pesquero, 1985. "Taller nacional fenomeno El Niño 1982-83. Investigacion pesquera" (Numero especial), Santiago,Chile. 254 pp. This is a collection of papers focusing on ENSO applications in fisheries in Latin America.

la Cock, G.D., 1986. 'The southern oscillation, environmental anomalies, and mortality of two southern African seabirds'. "Climate change", 8: 173-184

Lagos,P. and J. Buizer, 1992. 'El Niño and Peru: a nation's response to interannual climate variability'. In "Natural and technologies disasters: causes, effects and preventive measures", Chapter 17, 224-239.

Matarira, C.H., and L.S. Unganai, 1994. "A rainfall prediction model for Southern Africa based on the Southern Oscillation Phenomena". SADC/FAO Early Warning System, GCPS/RAF/270/DEN, Harare, Zimbabwe. 42 pages.

Pearce,F., 1994. 'Fire and flood greet El Niño's third year.' "New Scientist", 1908:9

Seleshi,Y., and G.R. Demarée, 1995. 'Rainfall variability in the Ethiopian and Eritrean highlands and its links with the Southern Oscillation' Index. J. Biogeogr., 22: 945-952.

Simard, A.J., D.A.Haines and W.A. Main, 1985. 'Relations between El Niño/Southern oscillation Anomalies and wildland fire activity in the United States'. "Agric. Forest Meteorol.", 36(2): 93-104.