392 



Fishery Bulletin 104(3) 



may reduce recruitment to natural reefs by functioning 

 as catchments for pelagic juvenile rockfishes. 



To investigate the possibility that an oil platform may 

 reduce recruitment of rockfishes to natural habitat, 

 we simulated drift pathways (hereafter referred to as 

 "trajectories") from an existing platform to nearshore 

 habitat using current measurements from high-fre- 

 quency (HF) radars. 



Materials and methods 



Species modeled 



Because trajectories derived from HF radar approximate 

 transport pathways of near surface water parcels, we 

 chose to model the movements of pelagic juvenile bocac- 

 cio iSebastes paucispinis) that dwell near the surface 

 during their time in the plankton. This historically 

 important recreational and commercial fishing target 

 in central and southern California (Love et al., 2002), 

 is also among the shallowest dwelling juvenile fishes 

 (Lenarz et al., 1991; Ross and Larson, 2003). 



Off central and northern California, parturition for 

 bocaccio occurs from January to May and peaks in 

 February (Love et al., 2002). Off southern California, 

 the species has a reproductive season that spans all 

 year, but most larvae are released from October to July, 

 although January is the peak month. Juvenile bocaccio 

 recruit to inshore waters from February to August off 

 central California, although May through July is the 

 peak season (Love et al.. 2002). The trajectory simula- 

 tion period from May through August was chosen to 

 span this principal recruitment season. 



Bocaccio range from the Alaska Peninsula to central 

 Baja California, and adults are usually found over high 

 relief boulder fields and rocks in 50-250 m of water 

 (Love et al., 2002). The fish most often settle in rocky 

 habitat covered with various algae or in sandy zones 

 with eelgrass. Juvenile bocaccio are commonly found 

 in drifting kelp (Mitchell and Hunter, 1970; Boehlert, 

 1977) indicating that the fish recruit to natural habitat 

 encountered in offshore surface waters. For this analy- 

 sis, we assumed that waters from the shallow subtidal 

 to the 50-m isobath represented suitable habitat for ju- 

 venile recruits. This choice reflects the lack of informa- 

 tion about suitable habitat locations in our study area 

 and likely results in overestimates of the abundance of 

 such habitat. 



Annual scuba surveys and submersible surveys 

 (1995-2001) in the Santa Barbara Channel and Santa 

 Maria Basin regions showed that YOY bocaccio inhabit 

 the upper 35 m around one or more platforms for each 

 year surveyed. Platform Irene {34°36.62'N, 120°43.40'W; 

 bottom depth 73 m) was selected for analysis because 

 fish recruited to it each year from 1995-2001 (Love et 

 al., 2001) and it was the site of the highest density of 

 YOY bocaccio observed from submersible surveys dur- 

 ing these years (Love et al., 2003). Moreover, from May 

 through August, 1999 and 2002, Platform Irene was 



also in a region of good HF radar coverage, which al- 

 lowed computation of extensive trajectory ensembles. 



Ocean current measurement and trajectory calculation 



Near-surface ocean currents were measured hourly by 

 using an array of three HF radars (SeaSondes, manu- 

 factured by CODAR Ocean Sensors, Ltd. of Los Altos, 

 CA) operating at 12-13 MHz. At these frequencies, the 

 measurement is an average over the upper 1 m of the 

 water column (Stewart and Joy, 1974). The radars were 

 located at Pt. Sal, Pt. Arguello, and Pt. Conception (Fig. 

 lA). HF radars measure components of surface currents 

 by means of a Doppler technique, at spatial increments 

 of 1.5 km in range and 5° in azimuth from the radar 

 location. Surface current vectors in an east-north coordi- 

 nate system were computed on a 2-km grid by using the 

 least square technique described by Gurgel (1994). With 

 this technique, all current components obtained within 

 a 3-km radius around each grid point were combined 

 to estimate the surface current every hour. The 3-km 

 radius limits the spatial resolution of the near-surface 

 current fields. Emery et al. (2004) have described the 

 processing of the HF radar data in more detail. Fur- 

 ther discussion on the use of HF radars for measuring 

 near surface currents is given by Paduan and Rosenfeld 

 (1996) and Graber et al. (1997). 



Emery et al. (2004) assessed performance of the three 

 HF radars by comparing them with in situ current me- 

 ters at 5 m depth. They found that root-mean-square 

 differences in radial speed measurements between HF 

 radars and current meters ranged from 0.07 to 0.19 m/ 

 s. Recent observations comparing surface currents from 

 HF radars and drifters have indicated that differences 

 are substantially reduced if spatial variability in cur- 

 rent fields is accounted for (Ohlmann, 2005). 



The areas used for computing trajectories were off- 

 shore of Pt. Conception and Pt. Arguello as shown in 

 Fig. 2A for 1999 and Fig. 3A for 2002. These areas 

 were selected to maximize the spatial coverage, and to 

 minimize the inclusion of grid points with low temporal 

 coverage. Variable coverage from individual radars re- 

 sults in differences in coverage between years. Boundar- 

 ies of nominal coverage areas were oriented along and 

 perpendicular to isobaths. Platform Irene is about 2 km 

 from the inshore boundaries, which lie along the 50-m 

 isobath (Figs. 2A and 3A). At times actual radar cover- 

 age exceeded nominal coverage boundaries as may be 

 seen by comparing sample computed trajectories (black 

 lines, Fig. lA) with the 2002 boundary (gray closed 

 curve. Fig. lA). Coverage in 2000 and 2001 for May 

 through August was inadequate for producing trajectory 

 ensembles around Platform Irene. 



A new trajectory was begun at the location of Plat- 

 form Irene every 4 hours from 1 May through 31 August 

 for 1999 and 2002. Positions along the trajectories were 

 determined by integrating current vectors forward in 

 time using a fourth order Runge-Kutta algorithm. Tra- 

 jectories ended where they encountered spatial gaps or 

 where they reached the edge of radar coverage. 



