1.50 



a. 



2 



o 50- 



tr. 



3 



< 



-1.00 



JAN 



JUNE 



MONTHS 



JAN 



15 



.10 - 



.05 - 



E o 



< 



to 



-05 



JAN 



JUNE 



MONTHS 



JAN 



FISHERY BULLETIN: VOL. 75, NO. 4 



150 



1.00 



i — i — i — i — i — i — r 



-1.00 



1.50 I — I — L. 



JAN 



i i i i 



JUNE 



JAN 



MONTHS 



FIGURE 5. — Mean annual variations: A. Koko Head temperatures (re = 18, 36, 54, 72). B. Koko Head salinities (re = 18, 36, 54, 72). 



C. Christmas Island temperatures (re = 20, 40, 60, 80). 



of the year. Consequently, the mean annual curve 

 reflects the temperature trends evident in indi- 

 vidual years in that warming lasts between 1 and 

 2 mo longer than cooling and the cooling rate 

 is higher than the warming rate. 



Departures of the mean annual salinity varia- 

 tion from the annual sinusoid, evident in Fig- 

 ure 5B, are not significant. 



In contrast to the Koko Head spectra, the ampli- 

 tude of the 6-mo sinusoid at Christmas Island 

 is large enough to produce a significant modifica- 



tion of the annual sinusoid (Figure 5C). The abso- 

 lute amplitudes of the 12-, 6-, 4-, and 3-mo 

 sinusoids are 0.43°, 0.21°, 0.04°, and 0.003°C, 

 respectively. Thus, the mean annual temperature 

 variation at Christmas Island has the typical 

 interference pattern produced by a 12- and a 6-mo 

 sinusoid as illustrated in Figure 6. The residual 

 curve, namely the difference between the mean 

 annual curve and the annual sinusoid, is approx- 

 imately the 6-mo sinusoid. 



Residual Variations 



0.8 



0.6 



0.4 



0.2 



K 

 3 



< 



0. 



5 -0.2 



-0.4 



-0.6 



-0.8 



JAN MAR MAY JULY SEPT NOV JAN MAR 



MONTHS 

 FIGURE 6. — Interference patterns of sinusoids for mean annual 

 variation at Christmas Island. Solid line - n = 20, 40, 60, 80; 

 dashed line- annual sinusoid (n = 20); dotted line- remaining 

 variation (re =40, 60, 80). 



774 



The dominant feature in the short-term curves 

 (panel C of Figures 1, 2, 3) is the annual variation 

 superimposed upon which is the variability that 

 characterizes each year. This "residual" variabil- 

 ity is obtained by evaluating in Equation (2) the 

 short-term variability without the annual sinus- 

 oid and its harmonics (n = 19-35, 37-53, and 

 55-71 for Koko Head, and n = 21-39, 41-59, and 

 61-79 for Christmas Island). Residual variability 

 is shown in panel D of Figures 1, 2, and 3. 



The residual curves are the interference pattern 

 produced by all the sinusoids used in the evalua- 

 tion. The irregular amplitudes and periodicities 

 occurring at intervals of more than 1 yr give an 

 impression of amplitude and frequency modula- 

 tions. For example, in the Koko Head salinity 

 curve, relatively large perturbations occur in 

 groups during 1959, 1964-65, 1967-68, 1969-70, 

 and 1972-73. In the Christmas Island residual 

 temperature curve, relatively large perturbations 

 during 1955-60 are followed by smaller pertur- 

 bations during 1960-65 and by larger perturba- 



