FISHERY BULLETIN: VOL. 84, NO. 3 



time scales. Hence, it seems plausible that northern 

 anchovy spawning strategies have developed in 

 response to the relatively predictable seasonal and 

 spatial trends in the California Current. Possible 

 relationships between time and location of fish 

 spawning and the currents off the west coast of 

 North America have been discussed by Parrish et 

 al. (1981). They noted that in the Southern Califor- 

 nia Bight the Ekman (wind-driven) currents are 

 generally diminished relative to other areas along 

 the coast. This reduced offshore transport is favor- 

 able for the retention of fish eggs and larvae. 

 However, some weak offshore directed Ekman 

 transport is consistently present in the Southern 

 California Bight year round (Nelson 1977; Parrish 

 et al. 1981; Bakun and Parrish 1982). 



Smith (1972) analyzed historical records of north- 

 ern anchovy larval distribution in the Southern 

 California Bight and found that samples taken 

 farther offshore had a higher proportion of older lar- 

 vae than that of samples taken nearshore. Assum- 

 ing a uniform spatial and temporal distribution of 

 spawning, this result implied that a significant frac- 

 tion of northern anchovy eggs and larvae were 

 transported offshore after nearshore spawning. 

 Bailey (1981) found that the average distance off- 

 shore of Pacific hake, Merluccius productus, larvae 

 north of Point Conception was positively correlated 

 with offshore Ekman transport and that the magni- 

 tude of subsequent Pacific hake recruitment was 

 negatively correlated with offshore transport. 

 Hewitt and Methot (1982) compared the distribu- 

 tions of northern anchovy larvae sampled in 1978 

 and 1979 and found that the bulk of the larvae in 

 1979 were farther offshore than those in 1978 and 

 that mortality of 0-group northern anchovy was 

 greater in 1979 when compared with those spawned 

 in 1978. The year 1979 was one of enhanced upwell- 

 ing and colder temperatures (both concomitants of 

 offshore Ekman transport) relative to 1978. 



The studies cited above suggest drift may play an 

 important role in larval ecology, but the conclusions 

 drawn from plankton sampling must be viewed with 

 caution. Inferences drawn from field collections 

 about the drift of larvae usually carry the assump- 

 tion that both northern anchovy spawning and lar- 

 val mortality were uniform in space and time, 

 because the time and distance scales involved largely 

 preclude synoptic sampling of eggs and larvae 

 throughout the region. Hence, only correlative ex- 

 planations for the observed distribution can be 

 made, and other causal factors affecting the larval 

 distribution may be hidden. For example, an obser- 

 vation of greater proportions of older larvae in off- 



shore waters could also result from earlier spawn- 

 ing or greater early mortality (possibly coupled with 

 increased spawning activity) in those waters, and 

 not drift. Additionally, the mesoscale variability 

 present in the Southern California Bight and the 

 considerable patchiness of early and late larvae (due 

 to northern anchovy schooling behavior; Hewitt 

 1980, 1981a) further confound the conclusions 

 drawn from plankton samples and diminish the value 

 of interannual comparisons. Therefore, as an alter- 

 native to field studies, a simulation model of north- 

 ern anchovy drift in the California Current was 

 developed to help evaluate the role of drift in larval 

 ecology. The objective was to use the model to deter- 

 mine the effect of differences in northern anchovy 

 spawning location and time on the subsequent larval 

 distribution and to evaluate the effects on larval 

 distribution when offshore Ekman transport is in- 

 creased above its normal mean value. 



METHODS 



The drift simulation was based on the two-dimen- 

 sional (x,y) form of the advection-diffusion equation: 



dF d 



dt dx 



U-K,?l\+ 3 



vF 



dx 



dy 



where F = the concentration of eggs and larvae; 

 u and v = current velocities in the respective x 



and y directions; and 

 K x and K y = eddy diffusivity coefficients for the x 



and y directions. 



An analytical solution to this equation cannot be 

 evaluated relative to northern anchovy larval drift 

 in the California Current, although a numerical ap- 

 proximation that specifies larval concentration as 

 a function of location and time can be determined. 

 This was accomplished by approximating each of the 

 derivatives in the equation by weighted finite- 

 differences, so that the model was algebraically 

 formulated as the current and diffusivity-mediated 

 fluxes of larvae among geographic points in the 

 Southern California Bight. Apart from the assump- 

 tion that larvae continually maintained themselves 

 in surface waters, the northern anchovy were 

 assumed to be conservative and completely passive 

 drifters, i.e., no mortality or movement due to lar- 

 val swimming was incorporated into the model. 

 Details of the numerical methods used are presented 

 in Power (1984). 

 The geographic grid for the model was defined 



586 



