QUINN ET AL.: SOUTHERN OSCILLATION. EL NINO. AND INDONESIAN DROUGHTS 



loss of 3 mo time with each application of the 6-mo 

 running mean is a drawback to its use in forecast- 

 ing, so we also use the 3-mo running mean and 

 monthly plots of anomalies for locating inflection 

 points and evaluating trends on a more immediate 

 basis in support of forecasts. 



Anomaly trends for several indices were main- 

 tained in time section plots (Figure 2a, b) to 

 evaluate the Southern Oscillation and its expected 

 effects on the southeast trade system. Although 

 these limited records ( 25-30 yr) clearly showed the 

 close association of low indices with El Nino type 

 activity, and high indices with anti-El Nino condi- 

 tions (Quinn 1974, 1976), it was essential to 

 extend the study over a much longer period to 

 determine how frequently these climatic extremes 

 occurred 



The World Weather Records were searched for 

 the longest and most complete atmospheric pres- 

 sure records which could be used to extend our 

 study into the past. Madras, India (1841-1976); 

 Bombay, India (1847-1976); Djakarta, Indonesia 

 (1866-1974); and Darwin, Australia (1882-1976) 

 were within the area noted by Berlage ( 1957, 

 1966) to reflect Southern Oscillation-related pres- 

 sure changes in the Indonesian equatorial low 

 pressure cell. Santiago, Chile ( 1861-1976) had the 

 only long pressure record that could possibly rep- 

 resent Southern Oscillation-related pressure 

 changes affecting the South Pacific subtropical 

 high pressure cell. Although Santiago is generally 

 to the east of the subtropical high, it does reflect 

 these pressure changes (Berlage 1957, 1966). 



Correlations were run between the Tahiti- 

 Darwin index and the Santiago-Darwin index on 

 data for 1935-76 to further substantiate use of the 

 Santiago-Darwin index for representing the 

 Southern Oscillation and related El Nino type ac- 

 tivity. The Tahiti-Darwin index was used for this 

 comparison since it and the Santiago-Darwin 

 index showed similar amplitudes in their interan- 

 nual fluctuations. The similarity was due to the 

 fact that Tahiti and Santiago are separated by 

 analogous distances from the usual core of activity 

 in the subtropical high ( see fig. 10 in Berlage 1957, 

 or fig. 10 in Bjerknes 1969). At zero lag the correla- 

 tion coefficient between the two indices was 0.88. 

 The maximum correlation was 0.89 when the 

 Tahiti-Darwin index led the Santiago-Darwin 

 index by 1 mo. 



Figure 3a-h shows the triple 6-mo running 

 mean plots of pressure anomalies for Madras 

 (1841-1976). Bombay (1847-1976), Djakarta 



(1866-1974), and Darwin (1882-1976). They also 

 show similar plots of pressure index anomalies for 

 Santiago-Bombay ( 1861-81) and Santiago-Darwin 

 (1882-1976). The anomaly plots were used along 

 with other data in the evaluation of El Nino type 

 events reported over the past 135 yr. 



Classification of Events 



The classification of El Nino type events by in- 

 tensity is highly subjective since no two cases are 

 exactly alike with regard to time of onset, dura- 

 tion, areal extent, thermal departure, degree of 

 devastation, etc. Determinations concerning 

 event occurrence and intensity were primarily 

 based on: 1) reported disruptions of the anchoveta 

 fishery and marine bird life off the coast of Peru; 2) 

 scientific reports which discussed events that af- 

 fected the coastal regions of Peru and southern 

 Ecuador [e.g., Eguiguren ( 1894), Frijlinck ( 1925), 

 Murphy (1926), Hutchinson (1950), Sears (1954), 

 Schweigger (1961)1; 3) hydrological data for the 

 Peruvian coastal region; 4) sea-surface tempera- 

 ture data along the coasts of Peru and southern 

 Ecuador; 5) rainfall at coastal stations in Peru and 

 southern Ecuador; 6) height of preevent peaks and 

 depth of relaxation troughs in Southern Oscilla- 

 tion index trends; 7) related indications from 

 index component trends (when pressure compo- 

 nents from only one core of the Southern Oscilla- 

 tion were available); 8) sea-surface temperatures 

 over the equatorial Pacific; 9) rainfall data for 

 islands in the central and western equatorial 

 Pacific. 



We categorized events as strong, moderate, 

 weak, or very weak, depending on the intensity of 

 the activity and the time of year that it occurred. 

 The true El Nino sets in during the first half of the 

 year. A symptom which is common to El Nirios is 

 the presence of anomalously high sea-surface 

 temperatures off the coasts of southern Ecuador 

 and Peru. Other frequently mentioned features 

 include a southward coastal current, heavy rain- 

 fall, red tide (aguage), invasion by tropical nekton, 

 and mass mortality of various marine organisms 

 including guano birds, sometimes with sub- 

 sequent decomposition and release of hydrogen 

 sulfide (known as El Pintor) (Wooster 1960). 



Strong El Ninos are recognized as such by all 

 investigators; they involve positive sea-surface 

 temperature anomalies along the coast in excess of 

 3°C, they display most of the aforementioned fea- 

 tures, and the anchoveta fishery is seriously 



665 



