Species vary in the degree to which they exchange migrants among populations. This migration ('connectedness', or more precisely, "gene flow') is the cohesive force that maintains a species' integrity: those that frequently exchange migrants function as coherent units across a landscape, whereas those with little migration between populations function more as a collection of semi-independent units. The level of connectedness among populations is therefore a critical variable for effective management as well as for our fundamental understanding of the way species form, are maintained, and respond to selective forces in their environment.
Recent work has shown that even closely related species can have vastly different patterns of gene flow among populations, indicating that relatively subtle biotic and abiotic differences can have a disproportionate impact on genetic connectivity. We aim to determine the relative effects of different ecological factors on population genetic structure by comparing the genetic patterns of a large number of species. To this end, we in the process of building an original dataset describing the population genetics of a cross-section of intertidal species along the Pacific coast. We have also examined this question in acorn barnacles (Balanus glandula) and 11 species of rockfish.
Multispecies comparison of intertidal invertebrates
To date, we have compiled DNA sequence data for 21 common intertidal invertebrates, most of which have never before been studied with genetic methods, using a fragment of the cytochrome c oxidase I gene (COI) of the mitochondrion. Taken together, these genetic patterns provide a first look at the community genetics of the Pacific rocky intertidal.
Biotic variables influencing connectedness
Species with longer pelagic larval durations (PLDs) are commonly thought to travel further on ocean currents, linking distant populations by frequent migration events. For marine invertebrates that disperse primarily in the larval phase, PLD is therefore expected to correlate positively with gene flow. However, our growing dataset describes dramatic departures from this expectation. Two species of hermit crab (Pagurus hirsutiusculus and P. granosimanus) and a shore crab (Hemigrapsus nudus) exhibit highly restricted gene flow despite long PLDs of up to 70 days. Conversely, many species with shorter PLDs (7-10 days; e.g., Tegula funebralis, Lottia austrodigitalis and L. paradigitalis) show no signal of population subdivision. Closely related species with similar PLDs differ drastically in their connectedness (e.g., Pagurus hirsutiusculus vs. P. samuelis; Lottia digitalis vs. L. austrodigitalis), further undermining the link between larval duration and gene flow.
Instead, more subtle biological variables may drive the observed differences in connectedness: for example, many of the species with exceptionally low migration rates are crustaceans, whose larvae can exhibit behavior that leads them to remain close to their birthplace. Differences in spawning season can similarly influence the chances that currents will transport larvae. As we build our genetic database, we are also amassing biological information on each species to test for such links between biological traits and connectedness.
Abiotic variables influencing connectedness
Because invertebrates disperse mainly as planktonic larvae, abiotic variables such as oceanography and coastal geological features are expected to influence migration patterns among populations. Point Conception, for example, has long been thought to be a prominent barrier to gene flow because it represents a marked change in water current and temperature regime, and is coincident with the range endpoints of many species. However, it is already clear from the data in hand that oceanographic processes, such as coastal upwelling, may be more important in structuring marine populations. Pairwise population comparisons between Monterey and Santa Barbara populations show little evidence of restricted migration across Point Conception for most species in the dataset. By contrast, such comparisons between Monterey and Oregon or between Monterey and southeast Alaskan populations (which do not span similarly obvious geological features that might impede migration) show much larger genetic breaks. These results reinforce those from the rockfish and barnacle projects. Results from our rockfish project (see below) suggest that oceanographic upwelling events can restrict gene flow along the coast; species with larvae that settle during upwelling are more likely to have fragmented populations. It is possible that upwelling similarly affects many intertidal species, accounting for the limited gene flow between Monterey and Oregon seen within our invertebrate dataset. At present we are gathering more information on the larval ecology of our study species, as well as oceanographic data such as sea surface temperature and wind speed, in order to link these variables to the observed patterns of gene flow.
As we expand this dataset to encompass more species, geographic locations, and gene regions, we will be able to link organisms' ecological traits to their degree of connectivity to a degree never before possible. Our conclusions will have direct applications in marine reserve and threatened species management, and contribute significantly to our understanding of the way in which different ecological variables influence species' formation and maintenance of diversity.
From 2004-2006, we collected acorn barnacle (Balanus glandula) recruits at Hopkins Marine Station during their predominant settlement season in southern Monterey Bay (i.e. February-July). Over three years, the magnitude of recruitment alternated between periods of relatively low settlement (hereafter known as "valleys") and high settlement ("peaks"; Figure 3). Previous studies have clearly demonstrated that these recruitment patterns are tied to periods of active upwelling and relaxation of upwelling, respectively. Based on this correlation, we hypothesized that the scale of larval neighborhood might change over time. More specifically, that during periods of lower recruitment (i.e. active upwelling) the geographic extent of the larval neighborhood would be smaller and more local than during periods of higher recruitment (i.e. relaxation of upwelling), when we would expect the neighborhood to stretch further north.
By using information from a genetic cline along the California coast in this species, we were able to define the geographic boundaries for several peak or valley cohorts of recruits. With one exception, all tested cohorts had a larval neighborhood ranging from south of Pescadero State Beach to north of El Capitan State Beach. Although the genetic resolution of our markers limits our ability to set the southern limit of the larval neighborhood, large-scale oceanographic patterns and previous modeling work indicate that it is highly unlikely that B. glandula recruits settling in Monterey Bay are coming from the south. In the northern end, our markers have remarkably good resolution such that we know that a cohort that is genetically different from Pillar Point contains less than 20% of northern-type larvae. Indeed, our one exception, a peak in 2005, is due to a slight excess in northern individuals for that peak, which matches well with our initial prediction based on oceanography. Although the overall pattern of larval neighborhoods appears to be remarkably stable over time, it is hints such as this which encourage us to look more closely by focusing on determining individual larval trajectories. We are currently working to make this possible by expanding our set of genetic markers and following up this cohort analysis with individual genetic assignment tests. This work allows unprecedented insight into marine larval dispersal and the spatial and temporal patterns of population connectivity.
We compared population structure of eleven rockfish species sampled in similar fashion and analyzed with similar genetic tools to discern the relative importance of geography, adult habitat, speciation rates and other species traits to genetic structure along the west coast of the US. We designed a sampling scheme encompassing species with different adult habitat requirements and compared populations across the biogeographic break at Point Conception (south of Monterey) and across the ecological gradient seen north of Monterey. We included six species of nearshore demersal species as well as five species that occur in deeper waters, to examine potential differences that the two adult habitat choices might have on population structure. We include species in two subgenera with different speciation rates (Sebastosomus and Pteropodus) to asses the relationship between speciation rate and degree of structure. We use these data to ask the questions (i) Is there greater degree of phylogeographic structure northwards from Monterey Bay, where there is a known gradient in larval settlement as well as in oceanographic conditions, or south from Monterey Bay where there is a well documented biogeographic breakpoint and (ii) Is there a difference in patterns between species that have different life histories and habitat requirements or that have had different phylogenetic patterns?
Among the eleven species of the rockfish that we examined, we find a wide variety of clear genetic patterns. Two species show large genetic breaks in every test and at all loci. Three species show no differentiation. The rest show moderate to slight variation. Some genetic differences fall along the northern part of the coast whereas others fall to the south. Clearly, we find a wide variety of genetic and geographic patterns among these similar species.
Population differentiation to the north of Monterey, associated with ecological changes in oceanography and community ecology, appears at least as common as genetic differentiation across the biogeographic break at Point Conception to the south of Monterey. In addition, where we find genetic differentiation, its magnitude tends to be stronger to the north than to the south. For example, yellowtail rockfish shows a strong genetic difference between Oregon and Monterey Bay, while populations in Monterey Bay and the Santa Barbara channel are genetically less differentiated. Perhaps the most striking case of differentiation between Oregon and Monterey Bay is seen in blue rockfish. A weaker, but significant, genetic difference also distinguishes the Monterey Bay and Santa Barbara channel populations from each other. In addition, population structure among shallow water species of rockfish appears higher than among deeper species. This pattern, like the north-south pattern, is in accord with previous observations, but is not a rigid rule. This can be illustrated by two of the most differentiated species in our study: Blue rockfish is shallow but the yellowtail rockfish chooses moderate-to-deep habitats.
Phylogenetic analysis shows that there is no clear link between rapid speciation and strong population structure in Sebastes, such as there is in other taxa. Comparison of two subgenera suggests that larval duration is less predictive of gene flow than is larval settlement behavior. Among these similar species across the same coastal environment, we document a wide variety of patterns in gene flow, suggesting that interaction of individual species traits such as settlement behavior with environmental factors such as oceanographic currents can strongly impact population structure. The subgenus Sebastosomus contains five species sharing a common ancestor over 4 MYA, while Pteropodus contains nine species yet only dates back about 3 MYA. However, contrary to intuitive expectations, it is the former clade that shows strong genetic differentiation while the latter has little, if any, structure. These two groups also vary in reproductive characterstics: Sebastosomus settle during periods of upwelling, and have a 3-4 month pelagic larval duration (PLD), while Pteropodus settle during relaxation events, and have a shorter PLD. Again, contrary to expectations, the groups with longer PLD show more structure.
Of the five species in the subgenus Sebastosomus, only the widow rockfish, which occurs in the deepest waters among these species, shows no structure. The others include the species with the strongest structure we found: yellowtail and blue, and two species with moderate structure at mtDNA or microsatellites. Yellowtail rockfish are commonly found from Alaska to southern California whereas its sister species pair, black and olive rockfishes, tends to divide the latitudinal range: olives occurs most commonly between Cape Mendocino and Santa Barbara, while blacks are common from Alaska south to northern California, becoming increasingly uncommon further south. The part of the coast where these two sister species still overlap is the same part of the coast where the major biogeographic break in yellowtails is found. Likewise, the major geographic break in blue rockfish occurs at the area where olive rockfish give way to black. It is possible that olive and black rockfish represent a more advanced stage in the speciation process than seen within blue rockfish or within yellowtail rockfish, and that in each of these cases, the same geographic factors are at work to limit intraspecific gene flow and spark species differentiation.