Evaluation of fish biodiversity in estuaries using environmental DNA metabarcoding

yellowfin goby and Takifugu spp., can tolerate various environmental conditions [44, 45]. On the contrary, the endemic species C. takatsuensis was found only in a single habitat (i.e., the Takatsu River; S2 Table), confirming its known limited distribution [46]. This species is regis- tered as an endangered species on the IUCN red list together with seven other species found in this study (https://www.iucnredlist.org). It is of particular importance that the endangered spe- cies were successfully detected by the eDNA survey as it is a non-intrusive method for both the environment and the subjects [12]. Therefore, eDNA could be applied not only for biodiversity research, but also to detect rare, endangered species [33]. Also, tropical to subtropical species (e.g., Caranx ignobilis [47]; Spratelloides gracilis [48]), only occurred in the Sendai River (S2 Table), which is located at the southernmost sampling station of the study. These results were in accordance with the distribution of fish species. Environmental conditions and biodiversity Biodiversity is closely related to the environmental conditions [8]. The results of GLMs showed that salinity affected the proportion of freshwater and seawater fishes, which varied among the five rivers. Specifically, no freshwater species eDNA samples were detected in the Tama River, which can be explained by the sampling stations being near the coast and salinity being over 20 (Table 1; Fig 1C). The proportion of seawater species accounted for more than 50% at high

tide but decreased notably at low tide (Fig 2C). The Sendai River showed a very small propor- tion of freshwater species at the river mouth, which is relatively wide (>1 km), compared with Aka, Miya and Takatsu River (Figs 1 and 2F). It is believed that seawater easily enters into riv- ers with wide mouths, which causes freshwater from the river to disperse and dilute across the adjacent coastal areas. As a result, brackish and seawater species accounted for more than 90% of hits in this case. Besides the width of rivers, tidal range is another factor with a strong influence on species composition. The tidal ranges are very small in the Sea of Japan [49], ranging from 6 cm for the Aka River to 55 cm for the Takatsu River, on the day of the sampling (www.jma.go.jp). In contrast, the tidal range of the Tama, Miya, and Sendai Rivers, which are located on the Pacific coast, was 167 cm, 67 cm, and 227 cm, respectively. Not surprisingly, salinity and number of species differed between high and low tides in the Sendai River (Table 1; Fig 2F). In the Tama River, the number of species did not differ between high and low tides; however, seawater spe- cies decreased at low tide (Fig 2C). Species composition in the Aka River differed remarkably between the river mouth and coastal area; the proportion of freshwater species was about 30–40% at the river mouth but decreased to 8–10% in the coastal area, whereas seawater species increased from 18–19% at the river mouth to 42–60% in the coastal area. This pattern can also be explained by the width of

the river mouth, which is very narrow (ca. 100 m) and thus affects species composition (Figs 1B and 2B). A similar trend was observed for the Takatsu River, which also has a narrow river mouth (<300 m); freshwater species decreased and seawater species increased in the coastal area. The proportion of seawater species was especially small at the river mouth of the Takatsu River, where water sampled from the bridge located about 1 km away from the river mouth had a salinity of 0.1 at both high and low tides (Table 1; Figs 1E and 2E). In fact, GLM analysis revealed that salinity had a significant effect on the proportion of freshwater and seawater spe- cies (Table 2). Biodiversity was high at the river mouth of the Aka River, and in the coastal area of the Takatsu River (Fig 2B and 2E; S2 Table). As the number of species was almost identical at the river mouth of both rivers (34–37 species and 32–33 species, respectively), the observed change in biodiversity could be explained by two phenomena. First, as mentioned above, there are PLOS ONE Biodiversity of five estuaries PLOS ONE | https://doi.org/10.1371/journal.pone.0231127 October 6, 2020 10 / 15