In mid-1996 climate scientists who had been following the evolution of the La Niña event that was then in progress started a tentative and guarded dialog among themselves about the "next" El Niño. It was yet too early to see any clear signs of its likely onset, but the 1990s had been remarkably active thus far. At the time, neither the statistical seasonal forecast tools nor the more analytical computer models that were then being developed had track records sufficiently long to inspire much confidence in what they foretold. At the same time, these and other indicators were beginning to point towards the possible onset of El Niño within the next year.
Without doubt, the most remarkable thing about these informal discussions in the climate community was that they were being held at all. In a few short years climate forecasting had moved from reliance on a few poorly understood statistical tools to a quantitative science that probed and measured the oceans and atmosphere, modeled their behavior, and made predictions based on computer models of the physical system.
It is not an overstatement to say that these first computer models of projected El Niños ushered in a new era in climate prediction, with the promise of benefits that reach far beyond the laboratory, and indeed, around the world. Individuals, and farmers’ almanacs, and even weather bureaus have in the past ventured to release seasonal forecasts of the climate a year or more in advance. But never, ever, with the tools and analytical understanding that is needed to do it right. With this new era, we move closer to the realization of a longstanding dream: of seasonal climate predictions, made months or seasons in advance, that are sufficiently accurate to affect planning for food production, public health, and environmental risk reduction.
The Changing Face of Climatology
For much of its history the branch of atmospheric science known as climatology dealt chiefly with keeping track of the average daily, monthly and seasonal temperature and precipitation at any place. Defined in this way, the seasonal conditions in a given region summers in England, for example were taken to be a more-or-less constant. These means defined the "climate" and were not unimportant, for they served a number of practical purposes: from engineering design specifications to advice on when the home gardener should plan to put in the tomatoes.
Yet, as we all know, we rarely experience the "mean," and indeed what we remember most are the extremes. In truth, the mean weather at any place also changes from decade to decade and certainly from century to century. Most adults have sensed and commented that the winters or summers are different from what they remember from their younger years. While not all of these perceptions and stories hold up to close scientific scrutiny, there is now ample evidence that the climate changes continually. There are long-term changes of climate on geologic time scales, such as the coming and going of the major Ice Ages, but also intermediate and shorter-term changes as well. Until recently, most of the effort in documenting and understanding climate variability was aimed at explaining the very long-term changes in climate. In the last thirty years, however, climate researchers have been able to discover the course and some of the causes of more immediate climatic changes that wax and wane on time scales of years and seasons.
Interannual Climate Variability
The best documented of all climatic changes are those that describe year-to-year, or interannual changes in the climate. The reasons for the seasons spring, summer, autumn, and winter are thoroughly understood in terms of the 23 1/2° tilt of the Earth’s axis and its well-worn orbit around the Sun. But why should one winter be milder than another? Or one summer unusually hotter or cooler, or wetter or drier? There are a number of possible causes. But at least some of the explanation can be found in the connections that tie the tropical oceans to the global atmosphere.
Essential to understanding, modeling, and predicting the El Niño phenomenon was the realization that the oceans and the air were both involved, and linked together. The nature of these connections was described in the preceding article of this issue of CONSEQUENCES. Here we focus chiefly on the observed effects of the most recent El Niños and La Niñas on temperature and precipitation patterns, around the world.
Lessons From Past El Niño Events
The dual elements of El Niño the tropical ocean and the air above it allow climate scientists to monitor the progress of present or coming events by looking at indicators in both the ocean and the atmosphere. One of the most used is the averaged temperature of the surface water in the central equatorial Pacific, often called the NINO index. Another popular indicator called the Southern Oscillation Index, or SOI, is based on the air pressure at sea level in selected regions of the Pacific Ocean. Other, more complicated indices employ both oceanic and atmospheric observations.
These indices are generated and updated monthly, and are publicly available from many sources, one of which is www.cpc.ncep.noaa.gov In translating them, you need to know that El Niño episodes are usually characterized by positive sea surface temperature (NINO) indices and negative sea level pressure (SOI) indices. A companion phenomenon with opposite characteristics, known as La Niña, is associated with negative sea surface temperature (NINO) indices and positive values of the SOI.
A reason for compiling indices of this kind is to reconstruct a history of past events, and indeed, versions of NINO and SOI span a period of more than a hundred years. As with most historical data there are gaps in the record notably for ocean temperature during World War I and II but these account for a relatively minor fraction of the total record.
The availability of such records makes it possible to identify past occurrences of El Niño and La Niña, and a number of ways have been developed, based on different indices, to determine which years qualify as one or the other. No matter what is used, however, most studies agree that there were slightly more than twenty-five El Niño episodes over the past century, slight less than twenty-five La Niña episodes, and roughly fifty so-called "neutral" years, when equatorial Pacific waters were less disturbed (see Table 1).
This relatively large number of events has allowed investigators to define the mean, or average, conditions that prevail in various components of the climate system during El Niño and La Niña events. In early studies climatologists were most interested in understanding the conditions at the center of action: the sea surface conditions, surface winds, and pressure fields in the equatorial Pacific at the time of El Niño. It soon became evident that for practical applications there was far more interest in land temperature and rainfall patterns that were associated with it. As it turned out, in fact, rainfall patterns were much involved in the El Niño process. An example of the rainfall patterns that characterize El Niño is shown in Figure 1.
The definition of typical El Niño-related rainfall patterns provides the most basic, primitive, level of information. The patterns shown in Figure 1 can be expected to apply during most, although not all, El Niño events: in this case, the conditions that are shown prevailed in at least 80 percent of all El Niños in the past century. Thus, for example, when an El Niño is in progress, or forecast, we can expect greater than average rainfall over the U.S. gulf coast, accumulated over the October through April period, in about four of every five instances.
These were the general patterns that climate forecasters expected when the 1997-98 El Niño began to develop. While many of them appeared as expected, not all of them did. Nor should we have expected them to do so, given the probabilistic nature of the climate system and how these expected effects were arrived at. How the 1997-98 El Niño came upon the scene, and what actually happened to rainfall and temperature patterns, region by region, are summarized below.
The 1997-98 El Niño: Early Signs
In September 1996 a La Niña episode had passed its peak and climate forecasters and researchers began to look for signs of what the tropical Pacific had next in store. It was clear that the ocean part of the climate system was in position for another El Niño. A thick surface layer of very warm water in the far western Pacific, from the International Dateline westward to Indonesia, was being held in place by the force of the strong La Niña-related winds that swept across the ocean surface from east to west.
Some of the earliest theories of El Niño required these very ocean conditions with a large, warm pool of water in the west Pacific. Some of the good computer models depended on a more sophisticated version of the warm water pool that was tied to the total ocean heat content. In truth, as the historical record shows, these warm pool conditions in the western Pacific don't always produce an El Niño.
The outputs from an array of computer models and statistical tools in the early fall of 1997 were far from unanimous in their sea surface temperature forecasts for the next few seasons. As the autumn progressed, however, the computer model run by the U.S. National Weather Service (at its National Center for Environmental Prediction, NCEP) appeared to be faring best in predicting what was then happening to sea surface temperatures in the equatorial Pacific. Beginning in late 1996, their model had consistently predicted the onset of El Niño conditions in the summer or early fall of 1997. With each successive monthly run of the model, the forecast intensity of the episode increased and the time of onset moved ever earlier in the year.
Clearer forecasts; continued concern
Climate scientists, including those at the International Research Institute for Climate Prediction, monitored the evolving climate scenario with increasing anticipation as 1996 came to a close. Surface waters in the Pacific were beginning to warm at an alarming rate. By February of 1997 almost all vestiges of colder than normal water were gone from the equatorial Pacific, and in March patches of warmer than normal water appeared. El Niño had made its tentative entrance on the world stage.
The initial rate of surface warming in the equatorial Pacific was impressive, and it continued to accelerate throughout the spring of 1997. By June, climate scientists agreed not only that an El Niño was brewing, but that a significant and perhaps unprecedented warm episode was underway. And they were right.
The 1997-98 El Niño: Impacts on Regional Weather
Indonesia: Drought in the wettest part of the world
It did not take long for the atmospheric component of the El Niño to manifest itself. True to form, and almost in lock step with the rise in sea surface temperature, the westward-blowing trade winds collapsed, and the heavy convective rainfall that accompanies the warmest sea surface temperatures moved eastward. Indonesia and Borneo began to dry out almost immediately, with deficits in monthly rainfall in both March and April of 1997 of four inches and more, and record-small amounts in May and June. After July, most of Indonesia would not see near-average rainfall again until April of 1998 nearly a year of severe drought in what is normally the rainiest corner of the world.
Forty-five inches of rain had fallen on the islands of Indonesia from March through December of 1997. Almost four feet of water would be an impressive amount for many locations on the globe, but here it was only about half of what is normally expected during these months. These monthly rainfall deficits led to severe drought conditions, and by July 1977, brush fires began to spread out of control in Sumatra and Borneo. Many of these had been started intentionally, to clear more land for crops, since the drought had already decreased the expected yields per hectare. The smoke that towered upward from these fires became a health problem in many parts of the country, as well as a severe visibility hazard for both air and sea navigation that persisted for several months.
In the Celebes, the center of the Indonesian area that felt the brunt of this El Niño, there were two extended periods with virtually no rainfall: mid-May through the end of June, and early July through the end of September. During this 126-day period there were fewer than seven days with measurable rainfall, in a region where a quarter inch of rain is expected every day. For better than a third of the year this tropical rainforest experienced desert-like conditions.
At higher elevations in New Guinea, the effects of the drought on crops were compounded by a series of frosts. The end result was a crisis in malnutrition and food shortages for hundreds of thousands of Indonesians. These almost certainly contributed to the subsequent political unrest that eventually toppled the government of Indonesia the fourth most populous country in the world.
Australia: Dry, but the worst fears fail to materialize
El Niño typically exerts a major influence on the rain that falls in Australia, imposing drier than average conditions over most of a country that has extensive arid lands. The main rainy season in the northern sections of Australia runs from late December through April, with rainfall more evenly distributed throughout the year in the eastern and southeastern regions. With the onset of the 1997-98 El Niño, the rainfall season in Queensland and the Northern Territory was ended abruptly and about a month early, in March of 1997. This dramatic change gave rise to fears that Northern Australia was due for extreme drought in the next rainy season December 1997 through April 1998 and that continuous dry conditions would prevail throughout the remainder of the country for the duration of the El Niño.
During the 1982-83 El Niño, which was until then the largest in modern times, Australia suffered one of the worst droughts in its history. The drought culminated with dust storms and wild fires in southeastern Australia, causing severe damage and loss of life and property. The scars from those fires are evident today, after seventeen years, in parts of Victoria. Quite understandably, the early advent and apparent strength of the El Niño in early 1997 caused great concern over much of Australia that this was about to happen again.
Fortunately, the consequences of the subsequent drought and shifting patterns of rainfall were not as severe as most had feared. Virtually all of New South Wales and Victoria, both in southeastern Australia, experienced very much below average rainfall, and in some locations, record-low amounts. In Melbourne, the January through November period was its second driest in 140 years of records.
Yet, crop yields in Victoria were generally good. The few wild fires that occurred here and there were not as severe or damaging as in the 1982-83 El Niño. More than expected rainfall in the months of September and November 1997 was one reason for the lack of dire consequences, given the prolonged dry conditions. Another contributing factor was that the previous year, 1996, was extremely wet, so that reservoirs and soils were fully charged and more able to withstand a prolonged dry period. Dry conditions returned to southeastern Australia in December of 1997 and remained well into 1998, but the effects of this drought were relatively small, compared to what might have accompanied so potent an El Niño.
Northern Australia was also relatively dry from March through November 1997, which is a normally dry season in this part of the continent. The El Niño was still very much in evidence in December 1997, at what is traditionally the start of the rainy season. Extremely dry conditions prevail in this part of Australia in more than eight of every ten El Niño years, based on the past 100 years of record.
Queensland and Northern Territory braced for drought, and farmers there found difficulty obtaining loans for the upcoming growing seasons. Some culled their herds in anticipation of drought. Quite unexpectedly, however, the rainfall was above normal in both December 1997 and January 1998. What fell came primarily in a handful of episodic storms. These were not only sufficient to keep rainfall totals above average at many locations, but were also responsible for local flooding. These rainfall surprises, atypical for El Niño, are a reminder of the kind of work that remains to be done in unraveling and predicting the course of seasonal climate.
February and March 1998 were both considerably drier than average in northern Australia. In March only about half the expected rain fell in Northern Territory and Queensland, although earlier rains were sufficient to moderate the potential drought. Pockets of very dry conditions did indeed develop, but widespread drought never materialized. Overall, Australian crop yields for the 1997-1998 growing season were good.
The summer monsoon in India
Indonesia and Northern Australia lie near the heart of El Niño’s influence, and it cannot be surprising that rainfall in these regions is closely tied to the phenomenon. El Niño influences rainfall in areas that are far from its source, as was shown in Figure 1. Although the causes of many of these regional effects are now understood, they were first uncovered in the course of statistical studies by correlating known El Niño events with recorded rainfall, elsewhere in the world, without the guidance of a physical explanation. These purely statistical relationships are often called teleconnections.
As Stephen Zebiak recounts in the preceding article, much of the work on the El Niño was inspired by the turn-of-the-century attempts by Sir Gilbert Walker to improve forecasts of variations in the summertime, monsoon rains in India. He also made the first systematic study of the statistical correlations that linked Indian monsoon rainfall to a number of other meteorological variables. Among these was a relationship between monsoon rain in India and the global sea level pressure patterns that were later found to be an important element of El Niño.
El Niños are one of many predictors of summer monsoon rainfall, in the sense that at these times the monsoon rains fail to materialize. While not all droughts in India have occurred in El Niño years, almost all of the largest that are known, historically, do. The India Meteorological Department has been producing monsoon forecasts for several decades, based in part on the El Niño, although they also make use a number of other empirical indices such as one based on Himalayan snow cover, and other aspects of atmospheric circulation. In India in 1997, given the magnitude of the then-emerging El Niño, there was grave concern that the country was in store for a major drought during the coming June through September monsoon season.
The severe drought never materialized. India experienced near-average rainfall amounts during the 1997 monsoon season, although there were drier pockets in the south central peninsula. In June, July, and August the rainfall was well above average for the country as a whole. September was the only monsoon month that had less than average rainfall.
Some climate scientists argued that the timing of this El Niño beginning as it did in March, with rapid growth during the summer came too late in the year to have more than belated influence on the 1997 Indian monsoon season, thus possibly accounting for the dry conditions in September.
Floods and droughts in Africa
The rainy season in southeastern Africa generally lasts from October to April, in an area that includes portions of South Africa, Botswana, Zimbabwe, and Mozambique. The region, like India and Australia and Indonesia, generally experiences relatively dry conditions during El Niño years; but what happened in October 1997 through April 1998 was not so simple. October and November of 1997 were relatively wet in this corner of the African continent; December was quite dry; and with the exception of January 1998, dry conditions continued through the remainder of what is normally their rainy season. Although the rainy season, taken as a whole, was dry as expected for an El Niño the early wet start and episodic heavy rains in January were sufficient to limit effects of the drought on crops in many areas.
Nonetheless, there were some very dry pockets in southern Africa. In some areas of northeastern South Africa, there was no significant rain for almost three months, from early December 1997 through mid-February 1988. Farther to the north, in Zimbabwe, there were areas with virtually no rainfall from early February through the nominal end of the rainy season, in April.
Farther to the north, Eastern Equatorial Africa -a region that includes Kenya, southern Ethiopia, Somalia, Uganda, and Tanzania generally experiences more rainfall during El Niño years. There, the deluge associated with the 1997 El Niño was nearly unprecedented.
Rainfall patterns in this part of Africa are quite complex, with two rainy seasons in a typical year: March to May, and October to December. Generally, it's the second of these that experiences wetter conditions during El Niño.
The rain that fell in some places was five to ten times the October to December average. As much as fifty-six inches fell at some locations, and several others recorded thirty-five inches. In many of these areas only ten to twenty inches fall in a "normal" three-month rainy season. The far heavier rainfall in late 1997 resulted, not surprisingly, in severe flooding, infrastructure damage, and loss of life. The residual effects of the heavy rains are still being felt. In addition to the direct loss of life and property, an increase in vector-borne diseases has been reported, and chronic poverty and hunger have been made worse in much of the region.
The Indian Ocean
In general, the regions surrounding the Indian Ocean experienced the kinds of rainfall patterns in 1997 that are typically expected from El Niño. But as noted above, some of these changes in rainfall patterns in this part of the world were less severe and shorter lived than was anticipated. What happened? Preliminary analyses suggest that one of the reasons that rainfall was less affected in India, Australia, and southern Africa was the unusually warm sea surface temperature in the Indian Ocean.
The most surprising of the rainfall surprises were the near-average amounts that fell in northern Australia for the December through March period. The area of the western Pacific from Indonesia southward into northern Australia is at the heart of the regions where rainfall is most affected during El Niño years. While this proved true in Indonesia, much to the nation's detriment, Australia's widely expected dry conditions gave way to a series of heavy squalls and tropical storms. In much of the country there were only isolated pockets of dry conditions.
Subsequent research at the IRI and in Australia points to unusual patterns of sea surface temperature in the Indian Ocean as a likely cause of Australia’s "non-standard" El Niño conditions. Climate scientists have learned from this experience that a bigger El Niño is not necessarily better behaved, in terms of following expected patterns of behavior. As often seems to happen in meteorology and climatology, as soon as a phenomenon is thought to be well described and understood, it shows a different face. One of the most important lessons from the 1997-98 event is that even though El Niños can alter weather and seasonal climate on a global scale, their impacts on the local climate are much affected by conditions in the local oceans where they originate.
Important among these, in this most recent El Niño, was the surface temperature of the Indian Ocean, which during 1997 and 1998 was uncommonly warm. Throughout most of the preceding two decades, variations from season to season in the surface temperature of the Indian Ocean were very small. The averaged temperature rarely changed by more than 1°F, which is close to the estimated accuracy of the measurements. For this reason many scientists believed that studying the Indian Ocean Basin was less crucial to understanding global climate variability.
Then, in the fall of 1997, temperatures in a relatively large fraction of the Indian Ocean warmed 2 to 4° above the long-term average. A change of this amount had never been encountered in the modern era of satellite measurements of ocean temperatures. So vast a pool of unusually warm water is thought to have had sufficient influence on local wind and rainfall patterns to perturb the effects of the 1997-98 El Niño that was then in progress. These included the relatively "normal" summer monsoon rainfall in India, abundant November and December rains in northern Australia, and the devastating rains that fell in eastern Africa.
The first occurrences of unexpected heavy rainfall in eastern Africa led scientists at the IRI to look more closely at the significance of the Indian Ocean, and to add an empirically-based Indian Ocean Surface Temperature Forecast to the tools then used to make El Niño projections. As a result, rainfall forecasts for eastern Africa during the remainder of the event were more successful.
By the late autumn of 1997, temperatures in the Indian Ocean’s Bay of Bengal began to return to near average conditions, while in the Pacific El Niño conditions continued to prevail. As though to confirm the role that Indian Ocean temperatures had just played, the unusual rainfall patterns in the area came to an end, and the so-called "northwest," or "winter" monsoon brought ample rains to southern India and Sri Lanka. Conditions returned to those usually expected at times of El Niño.
Elsewhere in the tropics, rainfall patterns tended to follow the "typical" El Niño patterns from the fall of 1997 into the spring of 1998. In particular, extremely dry conditions characterized an area that stretched from south of India eastward to the vicinity of the international dateline. Drought plagued many of the small island nations in this region. West of the dateline, typically wet conditions spread eastward to the West Coast of South America.
El Niño and rainfall in South America
Typical hallmarks of El Niño in coastal Peru and Ecuador are extremely heavy rainfall and floods, coincident with the appearance of warmer than average water in the western equatorial oceans. The first warming of these waters was noted in April of 1997, but the magnitude of the change was too small to induce increased rainfall along the west coast of the continent until late in the year. Excessive rain began to fall in Ecuador and Peru in September, but the following month was again dry. In November and December the heavy rains returned, this time in earnest. In early 1998, Peru was beset with heavy floods and mudslides.
Excessive rain fell also in central Chile through much of 1997. Above average rainfall was recorded, commencing in April, for eight of the remaining nine months of that El Niño year. In June, twice the average amount fell over much of the area. Santiago, where less than about eleven inches are expected from May to October, recorded twenty-seven inches of rain fell in 1997. Heavy El Niño-related rain also fell in an area of southern South America, Northern Argentina, and Uruguay.
In contrast, northeastern and central Brazil were, characteristically for El Niño episodes, drier than average. There, the dry conditions were responsible for extreme shortfalls in crop harvests, and the attributed cause of considerable social unrest in the equatorial state of Ceará.
El Niño and rainfall in Central and North America
At mid-latitude locations, the effects of El Niño are normally not felt until the colder half of the year. Because the 1997-1998 event began in the spring and early summer, climate scientists in many countries had more than ample time to alert appropriate segments of their governments that the coming winter would likely be different in a number of predictable ways. For the U.S., the likely expectations were excess rainfall in autumn and winter months in the southern tier of states and along the West Coast, from central California to northern Mexico.
The heavy rains arrived in Florida and along the Gulf Coast as early as October 1997. Some areas in Florida experienced their wettest October to December in 104 years of record. Wetter conditions began as well in coastal regions of southern California, although widespread heavy rains and severe flooding did not materialize until early in 1998. In the first three months of 1998 there was heavy rainfall along the entire West Coast, near record snowfall in the Cascades and Sierra Nevada Mountains, and heavy snowfall in the Rockies.
At some places along the west coast of the country the rain was truly impressive. In San Francisco, more than thirty-nine inches fell in the period from November 1997 through April 1998, which is well over twice the average amount. Santa Barbara had over 50 inches, which is more three times what is normally expected. Along the southern tier of states, heavy precipitation picked up in intensity in early 1998, and continued at record or near-record levels, from Arizona through Florida. Florida’s Department of Forestry, which maintains a reserve of water pumps for fighting brush fires, was called upon to help bail out flooded areas.
The unusual strength of the El Niño was credited with contributing to record rainfall along the East Coast as far north as the New England states an area not generally thought to have been influenced by El Niño in most past episodes.
The situation farther to the south, in Mexico and Central America, was quite the opposite. The same El Niño-related meteorological conditions that brought so much rain to parts of North America were responsible for extremely dry conditions to the south. Mexico, in particular, starting in June of 1997 and continuing through July of 1998, had its driest year since 1945, and most of this unusual dryness can be ascribed to the El Niño. The Mexican drought was marked not only by its longevity, but also by its geographical extent, for there was drought in almost every corner of the country.
The March through May period is generally very dry over most of Mexico, but these months were extreme. Most areas recorded less than 20 percent of the average rainfall. Though the effects of El Niño have typically subsided, or ended altogether, by the spring that follows its peak, some unique features of this El Niño conspired to prolong the drought until June. The result was a much-delayed summer monsoon, and widespread forest fires that sent their smoke as far as the central United States.
The Demise of the El Niño and the Birght of La Niña, 1998
The end of the 1997-98 El Niño was eagerly awaited by many climate forecasters, who had been kept busy for so long in tracking and explaining the twists and turns of this remarkable episode. By the early spring of 1998, however, there were clear signs that there would be no intermission. This El Niño was not going to give way to more prosaic meteorological conditions, for as the curtain fell on the littered stage, Act I of La Niña was about to begin.
As with most other aspects of the ENSO cycle of 1997-98, the transition to La Niña conditions was in no way subtle or surreptitious. The sea surface temperatures that had remained at near record warmth through much of the year began to drop rapidly in late 1997. El Niño conditions continued to dominate the weather until at least March of 1998, in spite of the turnaround, since the levels from which sea surface temperatures were dropping were so much warmer than average. By July of 1998, the warmer than average surface temperatures in the equatorial Pacific had been replaced by colder than average values in all but the coastal regions of South America.
La Niña had begun. In response, over the next few months, the global pattern of winds adjusted to the effects of colder waters, and with these changes, rainfall and temperature patterns marched to a new beat.
The most dramatic shifts in global rainfall patterns were in the west Pacific, where unrelenting drought gave way to torrents of rain. The transition to La Niña didn't take long. In Indonesia, monthly rainfall increased from 15 percent below average in April to two times the average in June of 1998. As the year progressed, most of the rainfall patterns expected from La Niña fell into place. These included a wet summer monsoon in India, dry conditions in eastern Africa, somewhat wet in southern Africa, and wet in much of Australia.
Much of South America, on the other hand, did not feel a typical La Niña signal. The likely reason was that the cold equatorial waters usually associated with La Niña were largely confined to the open equatorial oceans, and did not reach, this time, the coasts of South America. These departures from typical La Niña patterns in the eastern Pacific tended to increase the precipitation in western South America, and in the process turned an expected dearth into above average rainfall.
The absence of La Niña-cooled waters off the South American coast also contributed to some significant drought patterns that were not anticipated. Chief among these was an extreme drought in the southwestern United States that extended into Mexico.
Dry conditions in the southern tier of the United States are expected during La Niña, but usually not until the autumn and winter. This time, as noted above, the drought took hold in the spring of 1998, and remained in force throughout the early summer. Drought diminished in the Southwest as the year progressed, but La Niña-related dry conditions characterized much of the southeastern United States in the fall and early winter of 1998. The dry pattern continued with reduced snowfall in the late winter of 1998-99, and as spring came in 1999, the southern Rockies were topped with only 75 percent of the average snowpack at many locations.
A climate surprise in early 1999 was the return of above average rainfall in Ecuador and Peru, with significant local flooding. In February and March, reports of heavy rainfall in the higher elevations of Peru and Ecuador in February and March had raised fears that these might be associated with a return to El Niño conditions. It was a false alarm. In this instance, enhanced Amazonian rainfall, characteristic of La Niña, appeared to have been displaced to the west, to fall instead on the western slopes of the Andes.
Summary: What Next?
The atmosphere and oceans took us through a remarkable sequence of La Niña and El Niño episodes in the closing years of the twentieth century. La Niña conditions in the summer of 1996 gave way, in 1997, to El Niño conditions that held on into early 1998, only to revert to another La Niña that continued through 1999. This alternating sequence of La Niña and El Niño events had not been experienced since the early 1970s. More remarkable and unlike the case some twenty-five years before each of them was forecast, well in advance, and the last two particularly well. The success of these auspicious advances in applied climatology can be attributed to an improved ocean observing system, and to advances in understanding the causes of El Niños and La Niñas, and the application of computer models to the coupled ocean-atmosphere system.
Capabilities for predicting the evolution of equatorial Pacific sea surface temperature are now good enough to allow climate forecasters to provide advance information and guidance to a wide range of potential users. More promising, still, are the first attempts to harness these climate models to predict seasonal temperature and precipitation in a more general sense. An example of one such precipitation forecast, issued by the IRI, is given in Figure 2, with a comparison of what in fact occurred.
This first generation of operational seasonal temperature and rainfall forecasts were issued in terms of the probability of lower, average, or above average conditions for the season involved. Such probabilistic forecasts represent today’s state of the art in seasonal climate prediction, and they will likely continue as the main product of seasonal prediction efforts. One of the major challenges that now faces the climate community is to build the connections with user communities around the world, so that such information can be put in the most helpful form.