Modeling larval dispersal and quantifying coastal connectivity based on a downscaling ocean model in Seto Inland Sea, Japan

Connectivity is defined as the probability that particles leaving a source patch (spawning site) arrive at a sink patch (settling ground) for a given advection time. It can measure stochastic processes of a wide range of larval dispersal due to chaotic coastal circulations (e.g., Mitarai et al., 2009). The present study aims at investigating larval networks in Seto Inland Sea (SIS), an about 450-km long, semi-enclosed, estuarine tidal channel connected to Kuroshio drifting off the two openings of SIS. We quantify coastal connectivity using Lagrangian PDFs of enormous amount of Lagrangian particles released in the modeled SIS circulation field in a double nested configuration based on ROMS at the horizontal grid spacing of 600 m (Uchiyama et al., 2012). The particles are released twice a day from 140 source patches with a radius of 5 km for 2 months in the winter 2011, and integrated over the advection time of 30 days, mimicking spawning and pelagic larvae of marbled sole.

The Lagrangian PDFs and the associated connectivity patterns are found to be generally heterogeneous. The connectivity matrices successively depict the larval networks in SIS. We separate SIS into 8 subregions (i.e., bays) to determine the networks among them. The fraction of the particles that exit to another region is about less than 40% except for the two entrance subregions, suggesting the intra-subregional transport is predominant. However, the particles are gradually transported eastward due to the residual mean clockwise circulation of SIS. We then evaluate the larval dispersal of marbled sole in a subregion, Harima Nada (HN). Destination strength shows that particles released in HN are mainly transported towards the northern shore of Shodo Island, and source strength indicates that the particles arriving there originate primarily from the western shore of HN. Those results suggest a potential utility of the ocean model-derived connectivity to quantify the larval networks in estuaries.