Patterns in connectivity and retention of simulated Tanner crab (Chionoecetes bairdi) larvae in the eastern Bering Sea

• High retention of modeled larvae in all major source regions.
• Strongest connectivity occurs in a northwestward orientation.
• Cross-shelf advection is very limited.
• Southeastern Bering Sea (Bristol Bay) receives very limited larval subsidies from other regions.

The eastern Bering Sea (EBS) population of Tanner crab (Chionoecetes bairdi) has exhibited high variability in recruitment to the commercially exploited stock since the late 1970s. Concurrently, apparent shifts in crab distribution have also been observed. Larval advection patterns and associated local retention offer a potential mechanism for these observations. The Regional Ocean Modeling System (ROMS) was used to simulate larval Tanner crab advection patterns over 1978–2004 based on larval hatching sites inferred from the distributions of reproductive females sampled during annual National Marine Fisheries Service trawl surveys. Connectivity among EBS subregions was examined by comparing start and end float locations after 60 days of simulated drift. High levels of retention (>50% of floats) were observed in the majority of source subregions, and contributed significantly to the total number of endpoints in each region. Patterns in advection and resultant interregional connectivity were variable, with strongest sustained connectivity occurring along shelf, within individual domains. Increased settlement potential in the outer domain and southern middle domain after 1990 is consistent with an observed geographic shift in fishery productivity. Apparent reliance of Bristol Bay on local larval retention validates recent spatial fishery management to conserve this area as a subpopulation.