FISHERY BULLETIN: VOL 76, NO. 3 



Wooster (1960), Idyll (1973), Miller and Laurs 

 ( 1975), and Caviedes ( 1975) furnished background 

 information on El Nino; Quinn (1974) discussed 

 monitoring and prediction; and Berlage (1957, 

 1966). Troup (1965), and Quinn (1971, 1976) pro- 

 vided background information on the Southern 

 Oscillation and how it relates to phenomena dis- 

 cussed in this paper. 



Definitions for terms frequently used in this 

 paper follow: The Southern Oscillation was origi- 

 nally identified by Walker (1924). It was loosely 

 defined by Berlage (1966) as a fluctuation in the 

 intensity of the intertropical general atmospheric 

 and hydrospheric circulation over the Indo-Pacific 

 region. The fluctuation is dominated by an ex- 

 change of air between the South Pacific subtropi- 

 cal high and the Indonesian equatorial low. 

 The differences in sea level atmospheric pressure 

 between sites representing the South Pacific sub- 

 tropical high and sites representing the Indone- 

 sian equatorial low are used as indices to repre- 

 sent the Southern Oscillation (Quinn 1974). 



The El Nino type event refers to the appearance 

 of anomalously warm sea surface temperatures 

 and abnormally heavy rainfall in the equatorial 

 Pacific and an invasion of anomalously warm sur- 

 face water off the coast of Peru and southern 

 Equador. This event, which is brought about by 

 relaxation from a prolonged period of strong 

 southeast trades, is represented by falling and low 

 Southern Oscillation indices (Quinn 1974). The 

 magnitude of the interannual relaxation and its 

 timing with relation to the regular seasonal relax- 

 ation (Southern Hemisphere summer) appear to 

 determine the strength of the El Nino invasion 

 along the Peruvian coast. Heavy central and west- 

 ern equatorial Pacific precipitation usually starts 

 a few or more months after El Nino initially sets 

 in, but this may not always be the case. By using 

 the term "El Nino type" we avoid arguments over 

 what is and what is not an El Nino and can then 

 account for events that evolve in a similar manner 

 but vary in timing, intensity, and extent. 



The anti-El Nino refers to the contrasting situa- 

 tion when a strengthening and strong southeast 

 trade system prevails (represented by rapidly ris- 

 ing and high Southern Oscillation indices). At 

 such times we can expect strong upwelling (due to 

 the divergent equatorial flow under the influence 

 of strong southeast trades and equatorial easter- 

 lies), anomalously low sea surface temperatures, 

 and abnormally low amounts of rainfall over the 

 equatorial Pacific. Also, off the coast of Peru, we 



find strong coastal upwelling, low sea surface 

 temperatures, lower than average sea level, and 

 generally favorable physical environmental con- 

 ditions for biological productivity (due to the up- 

 welling of nutrient-rich water from lower levels). 



METHODS 

 Data Processing 



Atmospheric pressure and much of the rainfall 

 data before 1961 were obtained from the World 

 Weather Records (Clayton 1927, 1934; Clayton 

 and Clayton 1947; U.S. Department of Commerce 

 1959, 1968). Data for 1961-76 were obtained from 

 Monthly Climatic Data for the World (U.S. De- 

 partment of Commerce 1961-76). We were primar- 

 ily interested in the large-scale interannual 

 changes. Therefore, we eliminated regular oscilla- 

 tions from the data, such as the diurnal cycle, by 

 using monthly mean values (or monthly amounts, 

 for rainfall), and the seasonal or annual cycle by 

 subtracting long-term average or normal monthly 

 values from the actual monthly values. Data so 

 processed show no particular regularity and no 

 apparent cycle (Panofsky and Brier 1965). The 

 filtered and unfiltered monthly anomalies were 

 used to detect, identify, and evaluate any unusual 

 changes that took place. 



Our interests were focused on fluctuations of an 

 intermediate scale (Southern Oscillation), with 

 periods ranging between about 1 and 6 yr. The 

 remaining short period fluctuations in the 

 anomalies were eliminated by filtering with a low 

 pass filter. At the other end of the time scale, there 

 m.ay be a gradual change of the variate over many 

 years which is part of oscillations that are long 

 compared with the record. These extremely long, 

 gradual changes were not a factor in our study. 



In earlier papers (e.g., Quinn 1974, 1976) the 

 12-mo running mean was applied directly to 

 monthly values of pressure, pressure differences 

 (indices), rainfall, etc. as a low pass filter. This 

 filter not only smoothed the data to some extent 

 but also eliminated the annual cycle. To more 

 clearly define the interannual fluctuations 

 (Southern Oscillation), we recently switched to 

 the use of the triple 6-mo running mean filter on 

 the monthly anomalies, which requires three suc- 

 cessive passes of the 6-mo running mean over the 

 data. It results in smoother plots and more clearly 

 defined peaks and troughs, which are of particular 

 assistance in establishing long-term trends. The 



664 



