A major challenge in marine ecology is describing patterns of larval dispersal and population connectivity, as well as their underlying processes. We used a biophysical model to simulate dispersal of eastern oyster, Crassostrea virginica, larvae and connectivity among a network of 10 no鈥恡ake reserves in a shallow, wind鈥恉riven estuary to assess the relative importance of spawning location, spawning date, larval behavior, larval mortality, and adult reproductive output to predicted dispersal and connectivity patterns. The location (i.e., natal reserve) and date of spawning relative to physical processes, particularly frequency of wind reversals, were the dominant drivers of dispersal and connectivity patterns. To a lesser extent, larval behavior (i.e., 3D vertical advection and ontogenetic depth regulation) and mortality modified dispersal and connectivity, whereas spatiotemporal variability in adult reproductive output was of minimal importance. Over a 21鈥恉ay larval duration, mean dispersal distance of passive surface particles ranged from 5 to 40聽km. Reserves were too small (1聽km2) relative to mean dispersal distances to promote extensive local retention (median 2%) and spaced too far apart (typically ~50聽km) to promote extensive inter鈥恟eserve connectivity (median 2%). Limited connectivity and local retention may preclude the network from being self鈥恠ustainable, thereby limiting its long鈥恡erm conservation and management benefits. In reserve systems characterized by limited connectivity, management efforts should focus on increasing connectivity by increasing the number or size of reserves to realize the benefits of improved adult demographics within reserves.