FISHERY BULLETIN: VOL. 84, NO. 3 



(Fig. 1). The top 50 m of the water column is the 

 predominant depth range of anchovy larvae 

 (Ahlstrom 1959), therefore mean anomaly values for 

 the surface and for 10, 20, 30, and 50 m of depth 

 were used in this study. The anomalies at each 

 standard depth were interpolated to model grid 

 nodes using the bivariate interpolation algorithm of 

 Akima (1978). Geostrophic current velocities normal 

 to each grid cell interface were computed for each 

 of the standard depths, and average geostrophic cur- 

 rent velocities were then calculated for a layer ex- 

 tending from the surface to 50 m. 



Wind speed and direction data used in this study 

 were from the data base summarized and discussed 

 by Nelson (1977). The raw wind observations were 

 converted to surface wind stress (t) values using the 

 relation 



where p 



C d 



w 



T = RC d W 2 



air density (1.22 kg m 

 drag coefficient; and 

 wind speed. 



3 ); 



The drag coefficient was computed as a function of 

 wind speed using the empirical relation of Amorocho 

 and DeVries (1980, 1981). The computed wind stress 

 vectors were partitioned by month of observation 

 and resolved into alongshore and cross-shore com- 

 ponents. A monthly mean wind stress component 

 for each model grid cell interface was then computed 

 by averaging the appropriate component of the 

 stress vectors in the 37 km by 37 km area bisected 

 by the grid cell interface. Total Ekman or wind- 

 driven transport in the direction 90° to the right of 

 the wind can be approximated by dividing the wind 

 stress by the Coriolis parameter (Neumann and Pier- 

 son 1965), and this calculation was performed for 

 the mean wind stress components. The mixed layer 

 depth in the California Current is seldom >50 m, 

 and is often <20 m in the Southern California Bight 

 during the summer (Husby and Nelson 1982). It was 

 assumed that Ekman transport occurring deeper 

 than 50 m was negligible, and the Ekman transport 

 values were converted to a mean wind-driven 

 velocity for the surface to 50 m layer by dividing 

 the transport by the 50 m layer thickness. 



The final current velocities were calculated as the 

 vector sum of the seasonal geostrophic and appro- 

 priate monthly Ekman components. Vector addition 

 of the two components appears to be a reasonable 

 assumption (Parrish et al. 1981), and no compensa- 

 tion for redistribution of mass owing to sustained 

 winds was performed. The final seasonal current 



fields for the simulations were January, March 

 (April geostrophic velocities plus March Ekman 

 velocities), July, and October currents. 



Figure 2 illustrates the general trends in the 

 California Current for the January and March 

 seasons. This figure should be interpreted with cau- 

 tion. Apart from the large potential differences 

 between actual synoptic conditions and the average 

 pattern used in the simulations, the resultant vec- 

 tor for a cell was necessarily computed for Figure 

 2 by averaging the current components of oppos- 

 ing cell faces and then calculating the resultant. A 

 distortion is introduced wherever components on op- 

 posite faces of a cell differ in magnitude or sign, so 

 that Figure 2 best represents features of the Califor- 

 nia Current that are consistent over several model 

 grid cells. The California Current is evident as two 

 regions of intensified southeasterly flow at the left 

 margins and midlines of the plots. During all parts 

 of the year except spring, the current turns toward 

 shore at the southern end of the Southern Califor- 

 nia Bight. A northwesterly flow near the coast sub- 

 sequently forms the inshore portion of a large 

 cyclonic eddy (the Southern California Eddy; Owen 

 1980) that occupies most of the Southern Califor- 

 nia Bight. During most of the year part of this eddy's 

 northeasterly flow continues past Point Conception, 

 to form the California Countercurrent (Hickey 1979; 

 Fig. 2, January plot). In the spring the southeast- 

 erly flow of the California Current moves closer to 

 shore to obliterate the surface portion of the 

 Countercurrent (Fig. 2, March). Tsuchiya (1980, fig. 

 2) gives a clear picture of the seasonal inshore- 

 offshore movements of the California Current at 

 CalCOFI lines 90 and 93. Close to shore in the south- 

 ern half of the modeled region there is another 

 region of intensified southeasterly flow, most evi- 

 dent in the March current plot. Lynn et al. (1981) 

 provided detailed illustrations of the geostrophic 

 flow regimes used in the simulations, and Nelson 

 (1977) presented graphical representations of the 

 wind stress fields along the west coast of North 

 America. Hickey (1979) presented a comprehensive 

 review of seasonal and spatial variations of the 

 California Current and the possible driving mech- 

 anisms involved, and Owen (1980) reviewed the in- 

 cidence and ecological consequences of eddies in the 

 California Current system. 



Two additional current fields were calculated in 

 order to assess the effects of increased offshore 

 directed Ekman transport on larval northern an- 

 chovy distribution. As mentioned earlier, the mean 

 wind stress is consistently directed downshore 

 during March in the modeled region, a condition 



588 



