FISHERY BULLETIN: VOL. 76, NO. 3 



Table 5. — Strong, moderate, weak, and very weak El Ninos 

 with intervals between events from onset to onset. 



Onset 

 year 



Onsel 

 year 



Years 



between 



onsets 



Onset 

 year 



Onset 

 year 



Years 



between 



onsets 



1864 



1868 



1871 



1873 



1875 



1877 



1880 



1884 



1887 



1891 



1896 



1899 



1902 



1905 



1911 



1914 



1917 



1918 



112 (cumulati 



average time 



1868 

 1871 

 1873 

 1875 

 1877 

 1880 

 1884 

 1887 

 1891 

 1896 

 1899 

 1902 

 1905 

 1911 

 1914 

 1917 

 1918 

 1923 



ive years between onsets) 

 nterval between onsets 



35 (number of intervals) 3 2 yr. 



first mentioned in Quinn and Zopf (1975).] In 

 some cases there was an initial fall from a large 

 preevent (primary) peak which was not fully in 

 phase with the seasonal relaxation (Southern 

 Hemisphere summer and early fall) and the result 

 was a relatively weak event; then, there was the 

 rise to a smaller secondary peak followed by relax- 

 ation to a secondary trough which was in phase 

 with the seasonal relaxation and resulted in a 

 stronger event. The length of time between the 

 two troughs was generally 18-22 mo and it is our 

 opinion that situations of this type may account 

 for many of the event-to-event intervals that fall 

 in the short 1-2 yr category. Examples of such 

 developments can be noted in 1950-53, 1962-65, 

 and 1973-76 (Figure 2a, b). Preevent peaks occur- 

 red in 1950, 1962, and late 1973-early 1974. The 

 first relaxation troughs following these peaks oc- 

 curred in late 1951, late 1963, and late 1974-early 

 1975, and weak or very weak events resulted in all 

 three cases. Then, there were rises to secondary 

 peaks by mid-1952, mid-1964, and late 1975, fol- 

 lowed by falls to troughs by early to mid-1953, 

 mid- 1965, and mid-1976, resulting in moderate El 

 Ninos for these latter years. We must be aware 

 that these situations can arise and should be par- 

 ticularly wary when a large preevent peak is fol- 

 lowed prematurely by a weak or very weak event. 

 (One must not lose sight of the fact that these 

 interannual fluctuations in the index anomaly 

 trends were used to represent the interannual 

 fluctuations in southeast trade and equatorial 

 easterly strength as affected by the Southern Os- 

 cillation,) Figure 6 demonstrates the similarity of 

 the three two-stage developments discussed 



above; a particularly obvious index trend was 

 selected to represent each case. 



The index trend between late 1872 and 1877 

 indicates a possible three stage development (Fig- 

 ure 3b), with a weak event in 1873, a very weak 

 event in 1875, and a strong event in 1877 (Table 1 ), 

 It is noteworthy that Indonesian droughts, which 

 are usually associated with El Nino, occurred in 

 1873, 1875, and 1877 (Berlage 1957), 



The preevent index anomaly peak has been re- 

 ported to be a reliable indicator for subsequent El 

 Nino type activity, and our long index record sub- 

 stantiates this viewpoint. We compiled statistics 

 on the climb time from trough to peak and fall time 

 from peak to trough from our long index anomaly 

 record to provide some general guidance for event 

 predictions. Figure 7 shows the applicable statis- 

 tics. Events usually set in while the index is fall- 

 ing and prior to the index trough inflection point. 

 Therefore, the contents of Table 6 and Figure 8, 

 which pertain to time between index peak and 

 subsequent event onset, can be used to further 

 refine event predictions. We assumed a March 

 onset time for all cases in arriving at values in the 

 column headed "Peak to event onset" (Table 6), 

 This assumption was made since month of onset 

 was not available for most of the early cases, and a 

 study of recent cases showed onset times to range 

 from January to May, 



INDONESIAN DROUGHTS 



What happens over Indonesia relates to the 

 Southern Oscillation (Berlage 1957) and is, there- 

 fore, an integral part of the activity affecting the 



E-D 1949 

 JF-D 1961 

 R-D 1973 



3- 



1950 

 1962 

 1974 



1951 

 1963 

 1975 



1952 

 1964 

 1976 



1953 

 1965 

 1977 



if) — 

 in 



< CO 



Q. LlI 



_i 



Q. 



cr 



2- 



YEARS 



Figure 6. — Recent examples of two stage developments using 

 triple 6-mo running mean plots of: 1) Easter-Darwin (E-D) index 

 anomalies (1949-53); 2) Juan Fernandez-Darwin (JF-D) index 

 anomalies (1961-65); 3) Rapa-Darwin (R-D) index anomalies 



(1973-77). 



674 



