Fish larvae dispersal in the Western Indian Ocean and implications for marine spatial planning

Many marine reef fish species have two distinct life stages: a larval pelagic stage lasting a few weeks followed by a benthic stage after recruitment (Dufour, 1992; Shima, 2001; McCormick et al., 2002; Lecchini & Galzin, 2003; Irisson et al., 2004). Larvae dispersal is a vital process for species evolution, genetic mixing and overall adaptation of reef fish populations (Caley et al., 1996; Shulman, 1998; Planes, 2002). Integrating ecological connectivity patterns into marine ecosystem management is important (Roberts, 1997; Mora & Sale, 2002) especially in a global context of severe fish stocks depletion (Jackson et al., 2001) and growing degradation of coral reefs (Wilkinson, 2004) due to the combined impacts of human activities and climate change (Roberts, 2002; Hughes et al., 2003). Indeed, if a fish population is replenished by its own recruits (i.e. self recruitment), the population should be managed locally to maintain the genetic pool. This assumption of self-recruitment is dominant in the literature (Jones et al., 1999; Bay, 2000; Cowen et al., 2000; Fowler et al., 2000; Kingsford, 2000; Ochavillo et al., 2000; Jones et al., 2005; Almany et al., 2007). However if there is substantial exchange between geographically distinct populations, those species should be managed at a broader scale (Botsford et al., 2003), for instance through larger single marine protected areas (MPAs) (Claudet et al., 2008) or spatial reserve networks (Lockwood, 2002).

Ocean currents are the most influential parameter of larvae movement during the pelagic stage (Masterson et al., 1997) even if larvae have swimming and sensory capabilities that enable them to control part of their dispersal (Sale & Cowen, 1998; Kingsford et al., 2002; Leis, 2002). Determining larvae dispersal patterns based on in situ observation has always been a major challenge due to the small size of larvae and long dispersal distances (up to hundreds of kilometers from their initial release site, Leis, 1984; Victor, 1987; Clarke, 1995). In response, a variety of approaches have been developed to assess patterns of larvae dispersal and fish population connectivity across the marine environment, including genetics (Shulman, 1998; Planes, 2002), tagging (Jones et al., 1999; Swearer et al., 1999), stable isotopes (Peterson et al., 1985; Schwarcz et al., 1998; Blamart et al., 2002), otolith chemistry (Fowler et al., 1995; Campana et al., 1997) and otolith shape analysis (Smith, 1992; Torres et al., 2000; Pothin et al., 2006). Given the limited feasibility of the abovementioned methods across large regions, numerical transport models have been developed to infer pattern of larval dispersal (Schultz & Cowen, 1994; Roberts, 1997; Cowen et al., 2000; Treml et al., 2008; Mora et al., 2011). These models are increasingly being used worldwide for the design of MPAs (Planes et al., 2009), fisheries management (Gaines et al., 2010) and disaster management (e.g. oil spills, tsunamis, cyclones, Allison et al., 2003).

In this study we model connectivity patterns inferred between reef ecosystems of the Western Indian Ocean (WIO) using a hydrodynamic connectivity model (Treml et al., 2008) implemented in the Marine Geospatial Ecology Tools - MGET software - (Roberts et al., 2010). The fish dispersal model integrates historical altimetry data and pelagic larval duration (PLD). After Fish larvae dispersal in the Western Indian Ocean and implications for marine spatial planning presenting the study area, data and results, we further discuss model improvements and implications of this connectivity analysis in term of marine spatial management.