276 
Fishery Bulletin 116(3-4) 
Table 2 
Results of the analysis of indicator species for species of Sebastes identified 
either visually or on the basis of genetics for samples collected off Oregon and 
Washington during 2005-2008, by month or year in which samples were col¬ 
lected. Separate analyses were conducted for months and years. Results are 
combined here for ease of presentation. Only the species that had significant 
(P< 0.05) indicators for a particular month or year are shown. 
Indicator value P-value Month or year 
Visual identification 
S. helvomaculatus 
Genetic identification 
59.0/34.3 
0.0002/0.002 
August 2008 
S. pinniger 
38.2 
0.002 
May 
S. entomelas 
37.4 
0.003 
June 
S. flavidus 
33.3 
0.002 
June 
S. helvomaculatus 
22.7 
0.009 
2007 
S. babcocki 
16.7 
0.001 
2007 
proach allowed us to identify a total of 29 species in 
our samples, and to assign individuals to the WEVZ 
complex. Of those 29 species, 5 were identified only in 
the morphological collection, whereas only 12 appeared 
in the molecular data set. We could also examine an¬ 
nual patterns and seasonal patterns in the presence or 
absence of larval and juvenile stages of specific rockfish 
species for spring and summer and relate these to en¬ 
vironmental variables at local and regional scales (both 
cross- and along-shelf scales). 
With the publication of mitochondrial and nuclear 
sequences from 105 of the 110 then-described Sebastes 
species (Taylor et ah, 2004; Hyde et ah, 2007), it be¬ 
came possible to identify larval juveniles at any size. 
Sequence-based identification methods have been pre¬ 
viously used to identify morphologically unidentifiable 
rockfish larvae and to relate distributions of these lar¬ 
vae to oceanographic features and natal habitat (Taylor 
et al., 2004; Hitchman et ah, 2012; Thompson et al., 
2016; Thompson et al., 2017), to describe the early de¬ 
velopment of larval rockfishes (Watson et ah, 2016), and 
to describe the early life history among lesser-known 
species (Yu et ah, 2015). As with previously published 
research in Southern California utilizing cytochrome 
b sequencing, we identified 2509 rockfish from 24 dif¬ 
ferent taxa and the 4-species WEVZ complex (plus 25 
unknowns), and therefore adding this approach greatly 
increased the level of species identification and distri¬ 
bution details that could be made available. 
Given the difficulty or impossibility of visually iden¬ 
tifying the earliest life stages of larval rockfish, genetic 
approaches offer a way to explore the diversity of im¬ 
portant previously unidentifiable rockfish species, thus 
providing new information on the timing of pelagic lar¬ 
val and juvenile stages and distribution of these previ¬ 
ously unexamined species. For example, NMS and our 
observations from the indicator species analysis that the 
canary, darkblotched, and widow rockfish are more com¬ 
mon among spring samples and the rosethorn rockfish is 
more prevalent in summer, revealed a clear relationship 
between abundances of certain species and the time of 
year sampled. Although some of these patterns have 
been previously identified, we were able to expand our 
identifications to include a much greater diversity of 
species, including many of those that are unidentifiable 
by visual means. This newly described diversity was 
then related to environmental factors, such as ocean 
temperature, upwelling intensity, and productivity re¬ 
gime (PDO and NPGO indices), and season to explore 
temporal and spatial drivers of species diversity or mat¬ 
ing success (Taylor et al., 2004; Thompson et al., 2016; 
Thompson et al., 2017). Although a detailed seasonal 
and annual analysis of the abundance and distribution 
patterns of rockfish species is beyond the scope of this 
study, we did observe some months and years where in¬ 
dividual species showed greater prominence through the 
indicator species analysis. The years examined in this 
study showed substantial differences in their oceano¬ 
graphic conditions which may have greatly affected the 
cross-shelf and along-shore distributions of rockfishes 
(Brodeur et ah, 2006; Ralston and Stewart, 2013). 
Previous research on ichthyoplankton diversity and 
concentration in the eastern North Pacific has shown 
spatial variation at the local (Richardson et ah, 1980; 
Auth and Brodeur, 2006) and regional scale (Thomp¬ 
son et ah, 2014) and temporal variation over the short- 
and long-term (Brodeur et ah, 2008; Auth et ah, 2011; 
Thompson et ah, 2014), all of which are influenced by 
regional and basin-wide environmental fluctuations 
(Auth et ah, 2011; Auth and Brodeur, 2013). One uni¬ 
fying feature of much of this previous research is that 
some groups of larval fishes cannot be identified to 
the species level based on pigmentation or meristics. 
These unidentifiable groups include larval smelts (7 
species in the family Osmeridae), rockfishes (65 spe¬ 
cies), sanddabs (2 species of the genus Citharichthys), 
and snailfishes (17 species of the genus Liparis) (Rich¬ 
ardson et ah, 1980; Matarese et ah, 1989; Love et ah, 
