712 



Fishery Bulletin 93(4), 1995 



into seven geographical strata, with five to six stan- 

 dard stations located within each stratum. At each 

 station a 15-minute nighttime trawl sample was 

 taken at standard depth (30 m where possible, 10 m 

 at shallow stations). Following the cruise, identifi- 

 cation of rockfish specimens was confirmed in the 

 laboratory, standard lengths were measured, and a 

 subsample of otoliths were collected. 



Abundance indices for pelagic juvenile rockfish 

 were adjusted to account for interannual differences 

 in the size structure of the catch. After truncating 

 the data to include only the fully vulnerable portion 

 of the catch (i.e. SL >25 mm; Woodbury 3 ), additional 

 adjustments were performed in a two-step process. 

 Individual fish ages were predicted from standard 

 length [SL] measurements by using linear inverse 

 growth curves (age-f[Sh]) that were estimated for 

 each species during the 10-year period from 1983 to 

 1992. Specifically, the predicted age of species s in 

 yeary at standard length / is r . = cc sy + fi sy l, where 

 the a and /? were estimated by least-squares re- 

 gressions of age-length data gathered from micro- 

 scopic examination of otolith daily increments (see 

 Laidig et al., 1991; and Woodbury and Ralston, 1991 ). 

 If otolith data were unavailable in a particular year, 

 growth parameters were estimated from an analy- 

 sis of covariance of all the yearly data, by assuming 

 a common slope (daysmm -1 ) and the mean of 

 interannual intercepts. 



For each haul conducted and each species sampled 

 (subscripts not included), abundances offish of dif- 

 ferent ages were then adjusted to a common age by 

 using an exponential model with a constant mortal- 

 ity rate (Z), i.e. 



N; = A/ / exp[-Z(r*-r sW )], 



where Nf is the adjusted number of individuals of 

 length /, N { is the unadjusted number, and r is the 

 common age to which abundances were adjusted. In 

 all calculations x was set equal to 100 d, which is 

 generally representative of pelagic juvenile rockfish 

 ages during May-June (Woodbury and Ralston, 

 1991), and Z was fixed at 0.04 d" 1 . This latter figure 

 was based on combined estimates of mortality rate 

 for 1) larval shortbelly rockfish, S. jordani (Ralston 

 et al. 1 ); 2) settled juvenile blue rockfish (Adams and 

 Howard 4 ); 3) pelagic juvenile Pacific cod, Gadus 

 macrocephalus, and northern anchovy (Bradford, 



:( Woodbury, D. 1993. Natl. Mar. Fish. Serv., NOAA, 3150 Para- 

 dise Dr., Tiburon, CA 94920. Unpubl. data. 



4 Adams, P. B., and D. F. Howard. Natural mortality of blue 

 rockfish (Sebastes mystinus ) during their first year in nearshore 

 benthic habitats. Manuscript submitted to Fishery Bulletin. 



1992); and 4) pelagic juvenile Pacific whiting, 

 Merluccius productus (Hollowed, 1992). The A^* were 

 then summed over all lengths occurring within a 

 haul, yielding a haul-specific catch of each rockfish 

 species sampled that was adjusted for variability in 

 length composition. 



Final calculation of abundance statistics from our 

 midwater trawl surveys was based upon simple loga- 

 rithmic transformation of the data, i.e.y ■- = \og e [Xj k + 

 0.1], where x k is the length-adjusted catch taken in 

 haul j located in stratum k = 1 to 7. We estimated 

 the individual stratum means, variances, and stan- 

 dard errors for each sweep using conventional pro- 

 cedures appropriate to a stratified sampling design 

 (Cochran, 1977). The equally weighted stratified 

 mean was then used as a sweep-specific index of pe- 

 lagic juvenile abundance. Lastly, because the avail- 

 ability of pelagic juveniles to midwater trawling 

 shows marked seasonal change, the maximum value 

 of the stratified mean (among sweeps completed in a 

 year) was used to estimate relative annual abun- 

 dance, i.e. year-class strength. 



Direct underwater observation surveys 



Nearshore assessments were made by underwater 

 observers by using SCUBA at four locations on the 

 northern California coast (Fig. 1). Two of the study 

 sites, Dark Gulch (lat. 39°14'N; long. 123°46'W) and 

 Salmon Point (lat. 39°12'N; long. 123°46'W) in 

 Mendocino County, were monitored since 1983. In 

 1984, nearshore assessments were initiated at Horse- 

 shoe Point (lat. 38°36'N; long. 123°22'W) and at Fisk 

 Mill Cove (lat. 38°35'N; long. 123°21'W) in Sonoma 

 County, 100 km to the south. 



Each study site covers approximately 0.5 ha and 

 consists of high-relief rocky reefs surrounded by lower 

 reefs and boulders, interspersed by occasional sand 

 patches. Vertical water clarity was measured from 

 the boat with a white, plastic Secchi disk 20 cm in 

 diameter. Horizontal water clarity at a bottom depth 

 of 10 m was determined by estimating the distance 

 at which rock surfaces could be clearly observed. 

 Counts were not made when conditions were turbu- 

 lent, nor when visibility was less than 4 m. 



Observations for estimating year-class strength 

 began in late July, when settlement of pelagic juve- 

 niles was essentially complete, and continued 

 through the end of September. Young-of-the-year fish 

 were distinguished from older cohorts by their size 

 (40-50 mm SL in July), and from other species by 

 characteristic pigment patterns. 



Strip-transect counts were made between the 

 hours of 1000 and 1400 by observers using SCUBA 

 over bottom depths of 5-22 m. At each study site, 



