The need for an improved climate observing system was underscored during the planning stages of TOGA in the early 1980s, when the scientific community was caught completely off guard by the 1982–1983 El Niño, the strongest in over a hundred years. This El Niño was neither predicted nor even detected until several months after it had started. At the time, most in situ oceanographic data were available for analysis only months, or in some cases years, after they had been collected. So only a handful of scattered reports from islands and volunteer observing ships were available to track conditions in the equatorial Pacific in real time (delay of less than a day) or near-real time (delay of less than a month). Some SST reports were extraordinarily high and suggestive that an El Niño event might be underway. However, they were discounted as erroneous for several reasons. One reason was that there had been no "buildup" of sea level in the western Pacific by stronger than normal trade winds prior to 1982, presumed to be a necessary precursor of El Niño [Wyrtki, 1975]. Also, there had been no warming off the west coast of South America in early 1982, considered to be part of the normal sequence of events characterizing the evolution of El Niño [Rasmusson and Carpenter, 1982].
To complicate matters, in situ data were rejected from blended satellite/in situ SST analyses produced operationally by the U.S. National Centers for Environmental Prediction (NCEP), then known as the National Meteorological Center (NMC). These analyses indicated that the equatorial Pacific SSTs were near normal, or even slightly colder than normal, during much of 1982. However, the effect of the March–April 1982 eruptions of the Mexican volcano El Chichon on satellite SST retrievals was not fully appreciated at the time. These eruptions injected a cloud of aerosols into the lower stratosphere, where prevailing winds spread it around the globe at low latitudes within 3 weeks. The aerosols, whose effects were not included in algorithms to convert observed satellite radiances to SSTs, led to cold biases of several degrees centigrade in the satellite SST retrievals. Cloud detection algorithms interpreted these retrievals as cloud contaminated and replaced them with climatological mean SSTs. In situ data were then rejected because they differed so greatly from the satellite analyses, which were strongly biased toward climatology. It was only after reports from the R/V Conrad in September–October 1982 that the thermocline in the eastern equatorial Pacific was 50–100 m deeper than normal [Toole and Borges, 1984] that the scientific community realized to what extent existing data sources had misinformed and misled them and likewise how misguided was the notion of a "canonical" El Niño with a fixed pattern of stages from event to event.
Return to Discussion and Conclusion or go to Appendix B