Woodbury: Reduction of growth in otoliths of Sebastes entomelas and 5. flavidus 



683 



effect, and a model with interactions that esti- 

 mates a distinct otolith growth effect for each age- 

 by-year combination. These figures show that 

 although there are statistically significant inter- 

 actions, the signal for the year effect is not 

 strongly effected by excluding the interaction 

 terms from the model. The additional variance 

 explained by the model with interactions was 

 only 3.5% and 4.2%, for yellowtail and widow 

 rockfish, respectively. Because inclusion of an 

 interaction term contributes little to the explana- 

 tory power of the model, and analyses that in- 

 clude interaction terms are much more restric- 

 tive in terms of data requirements, the simple, 

 combined additive model was used, thereby al- 

 lowing the computation of a longer time series of 

 annual otolith growth effects (1980-87). 



Oceanographic databases 



Thirteen oceanographic variables were analyzed 

 to determine their relation with otolith growth 

 patterns. These included eight measures of ocean 

 temperature, three upwelling indices, a sea level 

 statistic, and the northeast Pacific atmospheric 

 pressure index. All ocean temperature data were 

 obtained from the Fleet Numerical Oceanogra- 

 phy Center, Master Oceanographic Observations 

 Data Set (Sharp and McClain, 1993). The data in- 

 cluded temperature at the sea surface, at 100 m, 

 at 200 m, and the depth of the 10°C isotherm. 

 Measurements were summarized from two 

 coastal areas (i.e. 36-39°N x 121-125°W and 39- 

 42°N X 123-127°W), which together encompass 

 the region where the fish were collected. Annual 

 means for 1980-87 were calculated from the 

 monthly means of October through September, 

 the otolith growth year as defined in this study. 

 Annual upwelling indices (Bakun, 1975) were 

 calculated from daily upwelling statistics. In this 

 instance, the mean of the February to April 

 monthly means was computed, as these three 

 months bracket the spring transition (Strub et 

 al., 1987) when nutrient-rich water is first up- 

 welled to the surface, initiating the biological 

 production cycle (Gushing, 1975). Upwelling data 

 were summarized from three localities (i.e. 36°N, 

 39°N, and 42°N), which represent the area offish 

 collections. Sea level data compiled at the San Fran- 

 cisco tidal station were used to compute monthly 

 anomalies, after correcting for differences in atmo- 

 spheric pressure (University of Hawaii^. Annual 



3.0 



2.6 



2 2 - 



£ 2.6 



2.2 



1.8 



Female 

 Bodega Bay 



Female 

 Eureka 



2.6 - 



2.2 



Male 

 Eureka 



Combined 

 Age1 

 Age 2 



— - Age 3 



82 



83 

 Year 



84 



Figure 3 



Comparison of mean annual otolith growth indices (1982-84) 

 among models with and without interaction terms for widow 

 rockfish. Bars represent ±1 standard error of the mean. 



' Sea Level Center, Univ. Hawaii, Honolulu, Hawaii. 

 Unpubl. data. 



1994. 



mean sea level anomalies were then calculated from 

 the monthly means of October through September 

 Lastly, the northeast Pacific atmospheric pressure 

 index is an indicator of atmospheric forcing upon 

 ocean waters (Beamish and Bouillon, 1993). This 

 index is the mean monthly difference in surface at- 

 mospheric pressure between 40°N, 120°W (Reno, 



