A bio-oceanographic filter to larval dispersal in a reef-building coral

Gene flow was shown to be limited between western and eastern Caribbean populations of the reef-building coral, Acropora palmata. However, some mixing was detected among populations near Puerto Rico. Our genetic analyses categorize A. palmata samples from the east coast of the Dominican Republic with the western Caribbean population, suggesting a filter to gene flow east of the Dominican Republic. To test the hypothesis of a present day bio-oceanographic filter occurring between Puerto Rico and the Dominican Republic (i.e., in the Mona Passage), we used a Lagrangian stochastic model (LSM) of larval dispersal, coupling coral life history characteristics with physical forcing. The model operated at two spatial scales: Caribbean-wide and focusing on the Mona Passage area. Results from the Caribbean-wide study showed no significant virtual larval exchange between the two populations. The small-scale model indicated that virtual larvae do not readily traverse the Mona Passage during the corals’ reproductive season. Larvae released from Mona Island, in the center of the passage, are retained in the lee within topographically steered eddies, which act, together with the larval competency period, as a de facto filter to dispersal. Combined, our findings reveal the location of a seasonal filter to gene flow and its mechanism. Understanding the causes of population differentiation is an ongoing effort in evolutionary biology. The oceanic environment provides few obvious physical barriers that may prevent genetic exchange of pelagic larvae between populations, so it is difficult to understand how populations differentiate. The numeric (how many?) and geographic (over what distance?) scale of larval movements has been at the center of a long-standing debate. Pelagic larval stages may connect geographically distant populations by riding with fast ocean surface currents (Heck and Mccoy 1978; Veron 1995). This view has been challenged by recent reports of localized recruitment and strongly structured populations (Thorrold et al. 2002). Biological factors such as reproductive strategy (Hohenlohe 2004), larval and/or recruit mortality (Schmidt and Rand 1999), and larval behavior (Carlon and Olson 1993; Altieri 2003; Fuchs et al. 2004) may interact with physical factors such as ocean currents (Alexander and Roughgarden 1996; Gaylord and Gaines 2000; Paris and Cowen 2004), eddies (often seasonal), and fragmented adult habitat (Johnson and Black 1995; Riginos and Nachman 2001) to restrict larval dispersal distance and influence its variance (Jackson 1986; Hohenlohe 2004). Understanding of the mechanism of such biophysical barriers to larval dispersal is the focus of this study. Boundaries between biogeographic regions as described by multispecies distribution records are logical places to also look for present-day barriers to gene flow. A correspondence cannot be assumed a priori because historical discontinuities rather than restricted larval dispersal at the location of the faunal break may underlie the observed species distribution patterns (Hellberg 1998). Point Conception on the California coast has received much attention because a biogeographic boundary coincides there with an ocean current convergence zone that is predicted to restrict along-shore larval dispersal (Dawson 2001). However, evidence for restricted gene flow at Point Conception from genetic studies is equivocal (reviewed in Burton 1998). Hohenlohe (2004) simulated the distribution of genotypes after dispersal by incorporating life history and oceanographic data from the Point Conception area into a deterministic simulation model and found that seasonal variations in ocean currents result in a leaky (i.e., 1 To whom correspondence should be addressed. Present address: University of Hawaii, Hawaii Institute of Marine Biology (HIMB), Coconut Island, P.O. Box 1346, Kane’ohe, Hawaii ([email protected]). Acknowledgments We thank R. Torres, J. Keehl, R. Albright, and E. Pugibet for providing the samples from the Dominican Republic. We also thank C. Rogers, E. Muller, and E. Mueller for providing samples from the U.S. Virgin Islands and the Bahamas, respectively, and A. Ortiz and E. Weil who sampled in Puerto Rico. Financial support for I.B. came from the NOAA-Fisheries Coral Reef Initiative through the Cooperative Institute for Marine and Atmospheric Studies (CIMAS), the Punta Cana Ecological Foundation, and the Friends of the U.S. Virgin Islands. We thank M.W. Miller for financial and logistical support as well as guidance and encouragement. M. Taylor, M. Hellberg, and R. Toonen provided valuable comments on the manuscript. Financial support for C.B.P. comes from R.K. Cowen, Maytag Chair of Ichthyology at the Rosenstiel School of Marine and Atmospheric Science. L.M.C. is supported by NSF grant OCE 03-271808. Funding was also provided by K. Lindeman from the Environmental Defense for the purchase of a 20-nodes PC cluster used to run the biophysical model. Limnol. Oceanogr., 51(5), 2006, 1969–1981 E 2006, by the American Society of Limnology and Oceanography, Inc.