UC Berkeley

This study was undertaken to identify population structure in the three rockfish species of the subgenus Rosicola through genetic analysis of six microsatelite loci applied to individuals sampled from 13 locations across the range of each species from Vancouver, British Columbia and San Martin Island, Mexico. Sampling each species throughout their respective ranges allowed for a more comprehensive and conclusive analysis of structure to inform proper stratification of stock assessments, catch tracking and management of these stocks. In addition, modeling potential evolutionary scenarios and testing for evolutionary processes shaping the current population structure were undertaken to better understand factors contributing to population structure and speciation in the subgenus. Results of this study were compared to previous research on members of this and other subgenera within the genus Sebastes to identify concordance in the phylogeography of subgenera or consistent processes between species indicative of common mechanisms contributing to genetic variation. Patterns observed were used to infer prioritization of future testing for population structure in the genus Sebastes and research to determine whether they are predictive.

No significant population structure was identified in the canary rockfish S. pinniger (Gill 1864) or the sunset rockfish S. crocotulus (Hyde et al. 2008). Three genetically distinct genetic clusters indicative of separate populations were identified in the vermilion rockfish S. miniatus (Jordan and Gilbert 1880) and population structure correlated with latitude and depth. While populations were found to overlap in their distribution to some extent, population structure was consistent with the boundary between the San Diegan and Oregonian biogeographic provinces near Point Conception, California. Additional population structure in the Southern California Bight south of Point Conception was correlated with adult depth distribution with a break around 60 m (30 fathoms (fm)). Assignment tests indicated that the young of year of each population as well as S. miniatus co-occur and reside in kelp forests before making ontogenetic migration to deeper depths with age. The deepest distributed of these clusters commonly occurs in depths of up to 100 m (50 fm), beyond which S. crocotulus was the most common species (Hyde et al. 2007). This population structure related to depth poses issues for stratification of historical data used in stock assessment and allocation of catch to each population as identifying characteristics have not been identified and management is already confounded by the presence of the recently identified cryptic species S. crocotulus.

Previous studies hypothesized that speciation between S. crocotulus and S. miniatus was the result of paedomorphsis in the form of concatenated migration to deeper depth isolating populations by depth arising from the greater area of nearshore habitat around the channel islands during early Pliocene glaciations when sea level falls as much as 300 ft. around 2.3 MYA (Hyde et al. 2008). Tests were conducted for repetition of such a pattern in forming population structure observed in S. miniatus. Timing of divergence in S. miniatus populations identified using assignment tests was estimated to be in the late Pleistocene only 102,429 (95% CI: 22,884 – 248,933) thousand years ago between clusters 1 and 2 and 255,141 (95% CI: 80,080 - 592,847) thousand years ago for their common ancestor and cluster 3. This is consistent with timing of glaciations, but results indicate northward expansion of a population, which may have occurred during an interglacial period. Our analyses indicate that mutation contributes significantly to population structure between species in the subgenus, but does not contribute to population structure in S. miniatus indicating that it is due to more recent drift or selection alone.

Population structure in S. miniatus consistent with northward migration of S. miniatus with lower genetic diversity indicative of a founder effect and private alleles accumulated in the time since divergence. No signal of bottleneck was present in any of the species or populations, though a model reflecting a reduction in population size in the northern population was found to have a high likelihood and posterior probability in Approximate Bayesian Computation analysis of potential evolutionary scenarios. It is possible that reductions in population size occurred too many generations ago to be detected using methods based on relationships between number of alleles and heterozygosity, given expansion of the population since its initial isolation. The results of modeling evolutionary processes indicated potential for gene flow between the northern population and deeper southern population potentially resulting in the shallow southern population. This may be the result of shared alleles from incomplete lineage sorting rather than admixture, or admixture occurred in the distant past as little evidence hybridization was observed in tests for individuals of hybrid classes. The southern shallow population had higher genetic diversity indicating that it may be an older stable population, in potential glacial refugia to the south. These results are also consistent with the possibility that population structure in S. miniatus as the result of northward expansion during the interglacial period.

Though the presence of a distinct deeper population in the southern California bight is consistent with paedomorphosis posed as the mechanism resulting in divergence between S. miniatus and S. crocotulus in Hyde et al. (2008), the processes leading to the division of adults on the basis of depth is still uncertain between the shallower and deeper southern population. We postulated that migration or shifts in abundance over time to glacial refugia in the south during glaciations may be associated with shifts to deeper depths to align with lower temperatures experienced at higher latitudes during the interglacial. Adults of each population would be potentially isolated from the shallower distributed progenitor taking advantage of ample nearshore habitat in the Southern California Bight during glaciations (Jacobs et al. 2007, Hyde et al. 2008). Having shifted abundance to the south and taken on adaptations to deeper depths, the now more deeply distributed population may remain isolated from the southerly shallow populations.

While concordance was imperfect, analogous structure in other Subgenera within the Sebastes including the Sebastosomas (Burford and Bernardi 2008), Pteropodus (Li et al. 2006) and Sebastomus (Rocha-Oliveras et al. 1999) that reflect population structure north and south of Point Conception as well as cryptic population structure with depth in cowcod (Hess et al. 2014), which we observed in the Rosicola. In other subgenera, differentiation is commonly accompanied by color variation and many of the sister species with similar indices of differentiation to populations identified in S. miniatus are identified as separate species due to their disparate coloration. This may in part be the result of sexual selection facilitated by internal fertilization (Gunderson and Vetter 2006).

Commonalities provide guidance on prioritizing future research to focus on species with distributions bridging Point Conception and a broad range of depth distributions. In addition, species with variable coloration should be considered for analysis as the line between species may be blurry but somewhat visible to the eye, needing only be confirmed by genetic analysis that can detect differentiation that in this case was indicative of a relatively recently diverged sister species (Frable et al. 2015, Narum et al. 2004). Nearshore species that release their larvae in the “sticky water” closer to shore that is slower moving due to turbulence and shear forces slowing its movement in the prevailing currents (Gunderson and Vetter 2006; Berntson and Moran 2009). This may contribute to maintenance of populations structure due to relatively inhibited larval transport compared to congeners in deeper waters spawning during periods of stronger flow of currents along the coast. These results provide guidance on selection of species for further analysis within the subgenus to identify what might be population structure or cryptic species that should be accounted for in stock assessments and management where possible. Nearshore species with large ranges bridging Point Conception, California, spawning in the spring during maximal upwelling, with broad depth range and with a large degree of variability in coloration are most likely to harbor genetic variation of interest to management and their analysis may provide further understanding the speciation process in this diverse genus.