• connectivity matrix;
  • conservation genetics;
  • dispersal;
  • ecological genetics;
  • evolutionary theory;
  • fish


  1. Top of page
  2. Abstract
  3. References

Understanding the scale of dispersal is an important consideration in the conservation and management of many species. However, in species in which the high-dispersal stage is characterized by tiny gametes or offspring, it may be difficult to estimate dispersal directly. This is the case for many marine species, whose pelagic larvae are dispersed by ocean currents by several days or weeks before beginning a benthic, more sedentary, adult stage. As consequence of the high-dispersal larval stage, many marine species have low genetic structure on large spatial scales (Waples 1998; Hellberg 2007). Despite the high capacity for dispersal, some tagging studies have found that a surprising number of larvae recruit into the population they were released from (self-recruitment). However, estimates of self-recruitment are not informative about mean dispersal between subpopulations. To what extent are limited dispersal estimates from tagging studies compatible with high potential for dispersal and low genetic structure? In this issue, a study on five species of coral reef fish used isolation by distance (IBD) between individuals to estimate mean dispersal distances (Puebla et al. 2012). They found that mean dispersal was unexpectedly small (<50 km), given relatively low IBD slopes and long pelagic durations. This study demonstrates how low genetic structure is compatible with limited dispersal in marine species. A comprehensive understanding of dispersal in marine species will involve integrating methods that estimate dispersal over different spatial and temporal scales. Genomic data may increase power to resolve these issues but must be applied carefully to this question.

Historically, it was assumed that marine populations were open or that larvae from all locations were mixed in the plankton, and as a result management regulations were applied over large areas (Lorenzen et al. 2010). But when direct tagging studies of larvae found self-recruitment (Jones et al. 1999; Swearer et al. 1999), the result was somewhat of a paradigm shift into thinking marine systems were more closed than previously thought (Mora & Sale 2002; Warner & Cowen 2002; Cowen & Sponaugle 2009). Most of the evidence for self-recruitment comes from coral reef fish, which are amendable to these types of studies because large numbers of larvae can be marked and released, most adults and larvae can be sampled for parentage analysis, and reefs are isolated in islands or embayments. However, the data from tagging studies are somewhat limited because they only estimate the amount of self-recruitment—one part of the entire ‘connectivity matrix’ between populations (Cowen & Sponaugle 2009). In addition, studies based on one year of data may not be representative because recruitment can vary widely on annual timescales (Cowen & Sponaugle 2009).

But large-scale genetic patterns in these species offer a alternative view of connectivity—low inline image implies that exchange of migrants is frequent on large spatial scales. However, as only one migrant per generation is enough to prevent large genetic differences from accumulating between populations in an island model (Slatkin 1987), it should not come as a surprise that inline image is low even if there is only moderate exchange among demes situated continuously along a coastline.

As a result of this stepping-stone model of dispersal, genetic distance between populations increases with geographic distance, a phenomenon known as isolation by distance (IBD) (Rousset 1997). IBD theory offers a middle-ground between short-term ecological estimates of connectivity and deep-time biogeographical patterns of genetic structure. The IBD slope (m) can be used to estimate the standard deviation of the dispersal distribution (σ) through the following relationship (Rousset 1997):

  • display math(eqn 1)

The above equation is valid for a one-dimensional habitat (when distance is not transformed) and estimates mean dispersal averaged over several recent generations. The IBD slope is less affected by rare dispersal events and only requires the distance between the samples rather than knowledge of the population structure (Rousset 1997). These properties make using Eqn 1 an appealing approach for estimating mean dispersal over multiple generations when information about effective density is available. For the design and evaluation of marine protected area networks (MPAs), mean dispersal over multiple generations is considered the critical metric because population build-up and persistence in a reserve depend on the average dispersal over multiple years, which generates multiple age classes (Gaines et al. 2010).

In this issue of Molecular Ecology, Puebla et al. (2012) indirectly estimate mean dispersal from IBD slopes in five species of coral reef fish over a 250 km range in the Mesoamerican Barrier Reef in Belize (Fig. 1). When estimating the IBD slope, genetic distance between populations inline image (Rousset 1997) or between individuals (Rousset 2000) may be used. Until now, other studies in marine populations have applied the former approach. But estimating inline image may not be ideal for continuous populations, because it requires deciding how to group samples into populations. The latter approach of measuring distance between individuals on a lattice, employed by Puebla et al. (2012), may more appropriate for continuously distributed marine species.


Figure 1. (A) Aerial view of the Mesoamerican Barrier Reef in Belize. (B) Bluehead wrasse (Thalassoma bifasciatum) are one of the species that exhibit a small, but positive, isolation-by-distance slope across the study area (photo credit: Edgardo Ochoa).

Download figure to PowerPoint

Using 11–13 microsatellite loci for each of the five species, they estimated the genetic distance between pairs of individuals, the effective population size (inline image) with approximate Bayesian computation, and the census density with scuba surveys along the transect. They found a positive IBD slope in all five species examined, although the slope was nonsignificant in four of the five species. The large scatter was expected because of the random nature of mutation and genetic drift; however, finding a positive slope in five independent species is unlikely by chance alone.

For the five species, they estimated mean dispersal between 9.4 and 52.5 km (confidence intervals of 1–142 km)— a range similar to direct estimates from tagging studies and that predicted by oceanographic models. These estimates are surprisingly small considering that some of the species they examined have long pelagic durations (13–94 days). To put the results in context, they performed coalescent simulations on a lattice to explore what kind of dispersal distributions would give patterns similar to the observed ones and analysed the simulated data with the same procedure as the microsatellite data. What they found was a pattern similar to their observed data: a small IBD slope that increased as the spatial scale of observation decreased and that was nonsignificant in many cases.

Their empirical data, in combination with the results from the simulations, provide a compelling story that mean dispersal in coral reef fishes is smaller than the potential for dispersal and that this is compatible with low genetic structure on large geographical scales. From an ecological perspective, such limited dispersal could have important consequences for the persistence of populations in MPA networks (Moffitt & White 2010) and the outcome of density-dependent regulatory processes (Hixon et al. 2002). However, from the perspective of evolutionary dynamics, the extent of gene flow will have important implications for whether marine species will be able to adapt to climate change: will advantageous alleles be swamped by migration, or will migration aid their spread throughout the range (Slatkin 1987; Lenormand 2002)? Unfortunately, most of the models that predict the dynamics of adapted alleles make assumptions that may be inappropriate for marine species that exhibit IBD and complicated patterns of connectivity. Determining the spatial distribution of dispersal will be an important component of predicting adaptation to climate change in marine species.

Although IBD theory is a useful way to infer mean dispersal averaged over a several recent generations, it has limitations. It assumes equilibrium between migration, mutation and drift, which is more likely to be met when the spatial scale of sampling is larger than the dispersal distance but not larger than 0.56σ/2μ (where μ is mutation rate; Slatkin 1993; Rousset 1997). An inherent difficulty in using IBD theory is the uncertainty around the parameters involved, especially effective density. It is generally unknown how different ways of estimating inline image are affected by gene flow from surrounding areas or unsampled populations, but some estimators break down under high gene flow that would be characteristic of marine species (i.e. Waples & England 2011). Also, IBD itself cannot infer what that underlying dispersal distribution looks like or whether it is asymmetric.

Ultimately, a comprehensive understanding of dispersal in marine systems will need to integrate theory with empirical oceanographic and genetic data, including estimates of dispersal at small spatial and short temporal scales (i.e. tagging or parentage analysis), estimates of mean dispersal averaged over several generations (i.e. IBD analysis) and larger patterns of population structure that have developed over long timescales (i.e. coalescent analysis). An important consideration in these analyses is that Euclidean distance between subpopulations can be a poor estimator of the connectivity between sites based on oceanographic distance (i.e. White et al. 2010). With the decreasing cost of genomic technology, the power to resolve these issues in nonmodel marine species is increasing rapidly and opens the door to a number of questions about which sampling strategies will be the most cost-effective (i.e. where should individuals be collected from and how many? how many single nucleotide polymorphisms are required to determine parentage or resolve fine-scale structure?). An important consideration of using genomic data to infer dispersal will be that the markers used are putatively neutral and not under selection or linked to a selected locus. Given that loci showing evidence of adaptive divergence have been found throughout the genome in model fish species (Jones et al. 2012), this may be a difficult criteria to meet. Nevertheless, additional detailed empirical studies that are combined with simulations and power analyses will be needed to yield a comprehensive picture of connectivity in marine systems.


  1. Top of page
  2. Abstract
  3. References

K.E.L. is interested in population genetics and evolution of high-gene flow organisms.