Abstract
Numerous studies have suggested that seagrass beds provide nursery habitats for juvenile fish in both tropical and subtropical regions. Most of these previous studies applied a landscape-scale perspective, in which seagrass beds and coral reefs are treated as independent, homogenous habitats. However, this perspective might overlook the microhabitat-scale perspective within the habitats, for example, the possibility that small-sized hard substrates (e.g., coral colonies) within seagrass beds might serve as fish nurseries. The present study aimed to examine the effects of the presence of microhabitats (small-sized coral colonies) within seagrass beds on the structure of fish assemblages. Fieldwork was conducted at Urasoko Bay, Ishigaki Island, Okinawa, Japan. Four habitat zones were selected: (1) seagrass bed with presence of massive coral colonies (SGCO), (2) seagrass bed without coral colonies (SG), (3) sandy bottom (without seagrass vegetation) with massive coral colonies (CO), and (4) sandy bottom without seagrass vegetation or coral colonies (SA). Six 20 m × 2 m line transects were established and monthly underwater observations were conducted between June and October in 2018 and 2019. A cluster analysis, analysis of similarity, and principal component analysis revealed that the fish assemblage at SGCO was significantly different from the other three habitat zones. This was because some fish species (e.g., Ostorhinchusishigakiensis and Lutjanus gibbus) were almost exclusively present at SGCO and rarely seen at CO, SG, and SA. Most individual fish belonging to these species were found on coral colonies at SGCO, suggesting that the co-occurrence of seagrass vegetation and coral colonies is essential for the habitats of these fish species. Although other fish species present at SGCO were also found at SG, three species, Parupeneus barberinus, Stethojulis strigiventer, and Lethrinus atkinsoni, were more abundant at SGCO with some found on coral colonies in this habitat zone. Several fish species that occurred at both SGCO and CO (e.g., Ostorhinchus properuptus, Cheilodispterus quinquelineatus, Chrysiptera cyanea, and Pomacentrus chrysurus) were more abundant or showed a greater size range at SGCO, suggesting greater survival rates in this habitat zone because of the co-occurrence of seagrass vegetation and coral colonies. This study demonstrated the existence of a unique fish assemblage structure at SGCO. Although the adoption of a landscape-scale perspective (three-dimensional structure of the vegetation) is necessary, a microhabitat-scale perspective that includes the presence of small hard substrates should also be considered to accurately evaluate the nursery function of seagrass beds.