Abstract
<p>The extent to which reefs are effectively connected to one another, and their potential to serve as sources of larval replenishment following disturbance, are topics of considerable interest in contemporary reef science. To date, most assessments of reef connectivity have emphasized long-distance horizontal dispersal of propagules from one shallow reef to another. The extent of short-distance vertical connectivity, however, has been largely unquantified. To fill this gap in knowledge, I developed DNA microsatellite loci for two Caribbean depth-generalist coral species with different life-history reproductive strategies (Montastraea cavernosa and Porites astreoides), and assessed connectivity in >1,200 coral samples collected from 3 depth zones (≤10 m, 15-20 m and ≥25 m) at sites in Florida (within the Upper Keys, Lower Keys and Dry Tortugas), Bermuda and the U.S. Virgin Islands (USVI). I also tested whether depth zonation in algal symbionts (Symbiodinium spp.) could limit effective vertical connectivity (Chapters 2 and 3). Finally, in Chapter 4, I led a collaborative seascape genetics effort to examine coral connectivity between the Flower Garden Banks (FGB) and the Florida Reef Tract at different depth intervals. This is a timely and important investigation because the FGB are located close to many oil and gas platforms in the Gulf of Mexico (GOM), including the Deepwater Horizon oil rig which exploded in 2010. Overall, Chapters 2 and 3 revealed significant genetic differentiation by depth in Florida (but not in Bermuda or the USVI) for both species, despite high levels of horizontal connectivity between all three geographic locations (M. cavernosa), or between Florida and the USVI (P. astreoides) at shallow depths. However, at all sites, and regardless of the extent of vertical connectivity, migration always occurred asymmetrically, with greater downward migration from shallow to deep habitats. Finally, whether or not M. cavernosa or P. astreoides exhibited depth zonation in algal symbionts did not appear to limit effective connectivity. Together, these findings suggest that: (1) depth is an important population structuring factor for corals, (2) the extent of vertical connectivity varies among and within geographic locations, likely as a consequence of local hydrology, (3) reproductive mode does not necessarily correlate with realized dispersal ability, and (4) shallow reefs are more likely to rely on distant (unimpacted) shallow reefs, rather than nearby deep reefs, to provide a viable source of new recruits following disturbance. Finally, Chapter 4 revealed high levels of gene flow between the FGB and the shallow Florida population of M. cavernosa, but not P. astreoides, suggesting limited gene flow among these regions. Results from biophysical modeling were in general agreement, suggesting that differences in reproductive mode and season might be important drivers of reef coral connectivity within the GOM region. Together, these findings suggest that FGB, despite its deep depth, might be an important larval source for shallow coral populations of broadcast spawning taxa in Florida. Furthermore, findings suggest that an oil spill originating in the GOM: (1) has the potential to impact coral communities in Florida by reducing recruitment from the FGB, (2) is more likely to affect broadcast spawning taxa like M. cavernosa, due to high levels of gene flow between FGB and Florida, and (3) regardless of coral reproductive mode, these impacts are more likely to affect shallow habitats, likely sinks for coral larvae produced at FGB. While deep coral populations in Florida may constitute refugia due to their partial isolation from the shallow population (see Chapters 2 and 3), they too might eventually be impacted if shallow populations were slow, or unable to recover.</p>