Williams and Ralston; Distribution and co-occurrence of Sebastidae off California and Oregon 



841 



43 



42 



41 



40 



2 39 



35 



34 



O o 



'®- ^ ° °^^r? «> ° ° ^ « O 

 ,. ... oogW" 



oO o O 



„ '^ ® o „ 0)0° °°„ =<»=^ 



b'^O 



> 



o 



oO o o 



3 . 



100 



200 300 



Depth (m) 



400 



500 



Figure 3 



Plot of total rockfish CPUE estimates by depth and latitude from the AFSC 

 triennial bottom trawl survey. Circle diameters are proportional to the 

 square root of total rockfish CPUE. Vertical bold lines represent an esti- 

 mate of the shelf break (see text). 



Escarpment. Likewise, the abrupt decrease in the amount 

 of shallow depth habitat from 35.5° to 36.5°N is directly 

 attributable to the occurrence of submarine canyons, and 

 the increase in habitat from 450 to 500 m at 35.0°N is due 

 to the Santa Lucia Bank (Fig. 2). The depth at which the 

 shelf break occurs seems to be fairly constant, generally 

 ranging from 100 to 175 m (Fig. 3). The relatively deep 

 (210 ml estimated shelf break at 36.5 N is due to the pres- 

 ence of Monterey Canyon (Fig. 3). 



For the combined Eureka, Monterey, and Conception IN- 

 PFC areas, the shelf trawl survey database pooled over the 

 1977-98 period totaled 1296 hauls that together captured 

 49 rockfish species, including an unidentified category (Ta- 

 ble 1 ). Based on the frequency of positive trawl samples for 

 each of the rockfish species, 26 were selected for detailed 



analysis (Table 1). In Table 1, the effect of changing survey 

 sampling objectives is evident in the interannual variation 

 in frequency of occurrence for some of the species. Perhaps 

 most noticeable is the abrupt decline in samples of aurora 

 (S. aurora ) and blackgill (S. melanostomus) rockfish in the 

 years 1980-92 (Table 1). Both are deep-water species and, 

 clearly, the relatively high frequencies of occurrence in 

 1977, 1995, and 1998 were due to increased sampling at 

 deeper depths that resulted from the altered objectives of 

 the sampling design discussed earlier 



The distribution of trawl sampling locations and CPUE 

 by depth and latitude indicated that the sampling pat- 

 tern followed some of the bathymetric features mentioned 

 above (Fig. 3). One notable feature was the paucity of 

 samples in the 36.5°N latitude region, which was partly 



