Introduction:
Coral reef ecosystems are under increasing threat due to the synergistic effects of habitat destruction, overfishing, eutrophication and climate change (Hughes et al., 2003, 2007; Hoegh-Guldberg et al., 2007). In response to these threats, management strategies that implement networks of Marine Protected Areas (MPAs) have gained momentum in the past few decades. Networks of MPAs can protect coral reef biodiversity from anthropogenic impacts either by eliminating the impacts of overfishing and habitat destruction, or by increasing ecosystem resilience to other anthropogenic disturbances (Russ and Zeller, 2003; McCook et al., 2010). For networks of MPAs to be effective they must meet three key elements. Individual MPAs must be (1) partially self-seeding (Almany et al., 2007, 2009), (2) adequately connected to other MPAs via dispersal (Jones et al., 2007; Almany et al., 2009), and (3) they must protect target organisms during life stages when they are most vulnerable to anthropogenic impact (Zeller and Russ, 1998). Accordingly, MPAs should be large enough to encompass individual home ranges of the target species and to ensure a portion of the larvae produced within a MPA settles within its boundaries (Almany et al., 2009). Furthermore, networks of MPAs must ensure genetic and demographic connectivity between protected areas. Connectivity is defined as the exchange of individuals between populations. Connectivity bolsters local resilience to stochastic demographic fluctuations and in so doing, minimizes genetic erosion, the risk of inbreeding depression and ultimately maximizes adaptive potential (Almany et al., 2009). Here we discuss how different life history strategies may affect the feasibility of achieving the three requirements for effective long-term conservation (selfseeding, connectivity, and protection). While sedentary organisms with a pelagic larval phase (most reef fishes and invertebrates), readily achieve this trinity (Planes et al., 2009), animals where dispersal only occurs as adults inevitably fail to meet all three requirements simultaneously (Figure 1). Here we propose a potential solution focusing on incorporating information on how habitat shapes adult dispersal to increase connectivity within networks of MPAs.
