Kope and Botsford Recruitment of Oncorhynchus tshawytscha in central California 



265 



differences, cf. Emery and Hamilton 1985, Mysak 

 1986). 



To test whether the first principle component was 

 actually related to ENSO events, we computed lagged 

 correlations between it and the Southern Oscillation 

 Index (SOI), a primary indicator of ENSO events (Table 

 4). The results show that the first principal component 

 is highly correlated with the Southern Oscillation Index 

 during the previous winter. Also, in all but one instance 

 (sea level height and spring SOI) the three oceano- 

 graphic variables are correlated with SOI in the pre- 

 vious winter in the way that would be expected from 

 their loading in the first principle component. 



To further simplify interpretation, we attempted to 

 use a single indicator of abundance. In correlations 

 computed thus far, no single population variable had 

 a demonstrably stronger relationship to environmental 

 variables tested. We therefore used the sum of com- 

 mercial catch, recreational catch, and spawner esti- 

 mates as a single indicator of annual population abun- 

 dance. For the corresponding deconvolved estimate we 

 use the deconvolution of this combination of these three 

 series, the deconvolution that was previously shown to 

 provide the best estimate of recruitment (Kope and 

 Botsford 1988). 



The results using the first principle component and 

 the total abundance estimate (Table 5) indicate an in- 

 fluence of ENSO-related conditions during the simimer 

 of the third year (a<0.05) and provide weak evidence 

 for an influence of ENSO-related conditions during the 

 first spring (a < 0.10). These relationships are reflected 

 in both the total abundance and the estimate of recruit- 

 ment to total abundance obtained through deconvolu- 

 tion. The estimate of total abundance is also correlated 

 with previously identified oceanographic variables. The 

 correlations with spring upwelling are - 0.478 (a< 0.10) 

 and -0.511 (a<0.01) for total abundance and decon- 



volved total abundance, respectively. The correlations 

 with summer ocean temperature are -0.551 (a<0.10) 

 and -0.540 (o<0.05) for total abundance and decon- 

 volved total abundance respectively. 



Discussion 



The negative correlation between abundance in the 

 third year and the first principle component of the 

 oceanographic variables is consistent with observed 

 effects of ENSO events on coastal oceanographic con- 

 ditions and fish populations. Productivity in the Califor- 

 nia Current, as represented by zooplankton biomass, 

 has long been known to be inversely correlated with 

 sea surface temperature (Reid et al. 1958, Reid 1962). 

 High productivity has traditionally been attributed to 

 upwelling, but this explanation was recently questioned 

 by Chelton et al. (1982) who proposed that higher zoo- 

 plankton biomass resulted from increased equatorward 

 flow in the California Current. The two proposed 

 mechanisms are difficult to separate because conditions 

 conducive to upwelling would also increase the strength 

 of the California Current. However, whether colder 

 water comes from the subarctic or from deeper coastal 

 waters, it is higher in nutrients than the warmer sur- 

 face waters, hence colder ocean temperatures in the 

 winter and spring are associated with higher produc- 

 tivity in the California current (Chelton et al. 1982). 

 Effects on productivity would be most noticeable dur- 

 ing strong ENSO events. For example, total catch and 

 average weight of chinook salmon landed in the com- 

 mercial and sport fisheries were lower during the re- 

 cent 1983 ENSO event (PFMC 1984, Pearcy et al. 

 1985, Johnson 1988). 



The weak correlation between marine environmen- 

 tal variables during the first summer and indices of 



