抄録
DeLury's method was expanded to 14 maximum likelihood models with variable catchability coefficients (qi) according to such environmental factors as water temperature, lunar cycle, and the intensity of ocean waves. The optimal model was selected by AIC, after the application of the models to catch-effort data from Japanese spiny lobster gilinet fishery. Confidence intervals of the estimators for each model were also discussed with a likelihood ratio test. AIC values greatly improved when variable catchability coefficients were assumed. Judging from the estimated parameter values, qi is large in conditions such as when (1) water temperature is high, (2) the phase of the moon is around the new moon, and (3) ocean waves are intense. When the relationship between qi and the water temperature was treated separately in two different functions for decreasing and increasing periods of water temperature, and when a five-day moving average of water temperature was introduced, the AIC values improved further. These variations in qi were probably attributable to changes in the activity of lobsters according to the fluctuation of environmental factors.
