272 
Fishery Bulletin 109(3) 
nity. Future studies of a possible link between middle 
and upper trophic level community structure and pe- 
lagic ocean habitats in this region could expand our 
efforts with the inclusion of additional hydrographic 
variables (e.g., dissolved oxygen, nitrates, silicates, tur- 
bidity, and frontal intensity). The relationship between 
zooplankton abundance and fish community structure 
is of particular interest. For example, cold-water cope- 
pod biomass has been shown to correlate strongly with 
survival of age-1 northern anchovy and may ultimately 
determine adult anchovy density (Litz et al., 2008). 
In the most comprehensive study of nekton com- 
munity structure and oceanography off Oregon and 
Washington to date, Brodeur et al. (2005) identified 
a larger set of water properties with significant cor- 
relations to fish assemblages. In their study, differ- 
ences in community structure were primarily related 
to seven environmental variables associated with dis- 
tance offshore and upwelling intensity: water depth 
and temperature, water transparency, chlorophyll con- 
centration, and three different macronutrients. The 
authors propose that future studies could develop a 
single metric combining multiple oceanographic vari- 
ables as a means of quantifying suitable habitats for 
pelagic species. Among the other surface trawl surveys 
in the CC during which environmental variables were 
measured, onshore-offshore differences in fish commu- 
nity structure were usually present, but north-south 
patterns were weak or absent. Several oceanographic 
features that vary with distance offshore (depth, tem- 
perature, salinity, and turbidity) were correlated with 
species distributions (Emmett et al., 2006). In a study 
describing the physical properties of biological hotspots 
in the coastal zone off southern Oregon, Reese and 
Brodeur (2006) found that distance offshore, water 
depth, temperature, density, and salinity accounted for 
the strongest correlations with species ordination axes, 
although the order of importance among these physical 
variables was different among years and seasons and 
chlorophyll was of little value in explaining community 
structure. Curiously, although the hotspots themselves 
persisted, the fish species inhabiting them did not. The 
authors found different sets of indicator species within 
each of two hotspots on four separate cruises. In all 
of these studies, an area farther offshore than that of 
the present study was sampled, usually at least to the 
shelf break. 
Our catch was dominated by a few highly abundant 
mid-trophic level species, as is typical of temperate 
upwelling zones worldwide. Examples include Engraulis 
mordax and Sardinops sagax in the California Cur- 
rent, E. ringens and S. sagax in the Peru Current, E. 
capensis and S. ocellatus in the Benguela Current, and 
E. encrasicholus and Sardina pilchardus in the Canary 
Current (Parrish et al., 1983). Chief among these “for- 
age” species in our study were northern anchovy and 
Pacific herring; lesser contributions were made by jack- 
smelt and Pacific sardine, depending on location and 
season. Northern anchovy composed more than half of 
the overall catch and were abundant in both regions 
but ranked seventh overall in frequency of occurrence, 
indicating large but scattered schools. Northern an- 
chovy abundance in surface trawls was also notably 
high and variable off Oregon (Brodeur et al., 2005; 
Litz et al., 2008), and larval anchovy were the most 
abundant fish in plankton samples from the Colum- 
bia River plume (Parnel et al., 2008). In midwater 
trawls off central California (Pt. Reyes to Pt. Concep- 
tion), Mais (1974) reported that 55% of hauls contained 
northern anchovy and greatest concentrations were in 
the south and <18.5 km offshore. In surface trawls in 
the northern CC, Brodeur et al. (2005) reported that 
Pacific herring, Pacific sardine, and northern anchovy 
together accounted for 76% of the catch, although their 
frequency of occurrence was relatively low. This pattern 
of clupeiform abundance was attributed to schooling 
behavior and patchy distribution. Subadult Chinook 
salmon were far less abundant in catches but were 
taken in a higher overall percentage of hauls, consis- 
tent with our findings and indicating a lower density 
and a more uniform distribution than for clupeids. In 
a broad survey of the CC between Vancouver Island 
and central California, 75% of 1.5 million fish taken 
in surface trawls were Pacific sardine and Pacific her- 
ring (Orsi et al., 2007). Pacific herring were positively 
associated with osmerids (true smelts) and juvenile 
salmonids, and together these three taxa (sometimes 
in combination with northern anchovy and other spe- 
cies) formed a distinct group typical of inshore habitats 
(Brodeur et al., 2004, 2005; Emmett et al., 2006; Orsi 
et al., 2007). In the present study, Pacific herring and 
juvenile Chinook salmon were co-dominant summer 
species in the NC region based on SIMPER analysis, 
but spatially herring were more dominant in the GF 
and juvenile Chinook salmon more dominant along 
the NC. Osmerids were far less abundant in our study 
than in those conducted farther north off Oregon and 
Washington and contributed little to the distinctions 
among communities in our area. 
Unlike herring and juvenile salmonids, Pacific sar- 
dines are usually grouped with an offshore migratory 
assemblage that includes Pacific mackerel ( Scomber 
japonicus ) and jack mackerel off Oregon (Brodeur et 
al., 2003; Reese and Brodeur, 2006). These three spe- 
cies appear to migrate from the southern CC farther 
north and onshore in unusually warm years (Brodeur 
et al., 2005; 2006). In spite of our limited cross-shelf 
coverage, the consistent importance of the variables 
DEP and DIS in matrix-matching tests strongly indi- 
cates that onshore-offshore gradients structure fish 
communities in our region as well, and implies that a 
distinct offshore assemblage perhaps similar to that of 
the northern CC exists in our area. However, in clas- 
sification analysis of selected species in our study (not 
shown), fishes typically associated with offshore habi- 
tats, such as Pacific sardine and jack mackerel, showed 
no clear relationship to each other or to distance off- 
shore, and Pacific mackerel were too infrequently 
caught to allow assessment of their distribution. The 
spatial coverage of ongoing trawl surveys conducted by 
