Chang et al. : The Antarctic Oscillation index as an environmental parameter for predicting catches of lllex argentinus 
205 
was relatively low after June; therefore, a correlation 
analysis of logt/ and environmental variables within 
the fishing season was undertaken for the time period 
from November of the previous year through June in 
the current fishing season. 
mental variables with a time lag of the previous year 
revealed that the logt/ was negatively correlated with 
subsurface seawater temperature in both the south 
and north locations in March (Table 1). The correla- 
tion analysis of lo gU and environmental variables that 
Time series of Antarctic Oscillation 
From 1984 through 2010, the monthly time 
series of the AAO index became positive 
approximately every 4-5 years, in 1985, 
1989, 1993, 1997-1999, 2001, 2004, and 
2008 (Fig. 5), and these positive phases 
had various durations. The autocorrelation 
analysis showed that the monthly AAO 
was correlated with the AAO of the previ- 
ous 3 months and was affected mainly by 
the value of the previous 1 month. 
Correlation between CPUE and environmen- 
tal variables 
The logt/ was negatively correlated with 
subsurface seawater temperature (at a 
depth of 5 m) at the south location from 
February through April, in the model and 
there was no time lag with the fishing 
season (Table 1). The logt/ was positively 
correlated with the AAO in December with 
no lag time during the fishing season. The 
correlation analysis of log U and environ- 
Lag (years) 
Figure 3 
Lagged autocorrelation functions (ACFs) for annual log-transformed 
catch per unit of effort (log U) of Argentine shortfin squid (lllex ar- 
gentinus) caught by Taiwanese jiggers from 1986 through 2010 in the 
southwest Atlantic. There was no statistical significance between all 
the time lags. The dashed lines indicate the 95% confidence intervals. 
