(JUINN ET AI. : SOUTHERN OSCILLATION. KL NINO. AND INDONESIAN DROL'OHTS 



affected (e.g., the 1957-58 and 1972-73 cases of re- 

 cent years). Moderate cases are recognized as El 

 Ninos by most investigators, and display typical 

 El Nino features to a lesser degree; maximum 

 monthly sea-surface temperature anomalies 

 along the coast usually peak in the 2.0°-3.5°C 

 temperature range (e.g., the 1953, 1965-66, and 

 1976-77 cases of recent years). The effects of a 

 moderate El Nino on the anchoveta fishery are 

 considerable, but less serious than for the strong 

 category. 



Weak events may or may not be recognized as El 

 Nirios by investigators; maximum monthly sea- 

 surface temperature anomalies along the coast 

 usually peak in the 1.0°-2.5°C temperature range, 

 but may appear relatively late in the year ( e.g., the 

 1951 and 1969 cases of recent years). Very weak 

 events are not considered to be El Nirios; 

 maximum sea-surface temperature anomalies, if 

 they penetrate into the coast, are in the 0°-2°C 

 range (e.g., the 1963 and 1975 events). The weak 

 and very weak categories are included in this dis- 

 cussion because the difference between weaker 

 and stronger events depends not only on the 

 height of the preevent index anomaly peak and the 



subsequent degree of relaxation reflected in the 

 southeast trade strength, but also on the timing of 

 this interannual relaxation. If the timing is in 

 phase with the regular annual relaxation (South- 

 ern Hemisphere summer and early fall), a moder- 

 ate or strong event is likely to occur; if they are out 

 of phase, a weak or very weak event is likely. 

 Relaxation troughs that occur near the end of the 

 year are usually associated with high Peruvian 

 coastal sea temperature anomalies in the latter 

 half of the year. The weak and very weak events 

 may not be of significance to the Peruvian an- 

 choveta fishery, but they do show up in the west- 

 ern equatorial Pacific rainfall and their larger 

 scale aspects may be significant from the 

 standpoint of associated global fluctuations. Fig- 

 ure 4 shows an example of how the recent events 

 were reflected in the Tarawa rainfall. 



The weaker events were included as EN(W) in 

 Figure 2a, b, since we have a fairly large amount of 

 evidence available from 1950 on. They were not 

 included in Figure 3a-h due to the decreasing 

 availability of evidence as we reach further back in 

 time. However, these weaker events, ascertain- 

 ed to the best of our ability from availAle data 



YEARS 



Figure 4. — Triple 6-mo running mean plot of anomalies of the difference in sea level atmospheric pressure (millibars) between Rapa 

 ( 27°37 'S, 144 °20 ' W) ( Austral Is. ) and Darwin ( 1 2°26 'S, 130°52 'E), Australia compared with a similarly filtered plot of Tarawa ( 0r2 1 'N, 

 172°55'E) (Gilbert Is.) rainfall anomalies (millimeters). 



671 



