Evaluation of fish biodiversity in estuaries using environmental DNA metabarcoding

fewer freshwater species in the coastal area of the Aka River. Second, marine biodiversity is higher in the Takatsu River because it is located in the southern part of Japan and in general biodiversity increases toward lower latitudes [50]. GLM results supported the increase in num- ber of seawater species when water temperature increased (Table 2). Composition and number of species were less straightforward for the Miya River, reflecting its complex geography and environment (Fig 1D). For example, the number of species was highest at the station 500 m along the left of the river mouth (Fig 2D), which can be explained by the junction of two rivers, the Miya River and the Tokita River. However, the number of species was lowest at the station 500 m to the right of the river mouth, where no freshwater species were detected; the reason for this was not clear. The narrow river mouth beside the sampling station (R1) might prevent the flow of freshwater to the right side of the coast, but salinity was lower on the right side than on the left side, and some freshwater species were detected at the station 1 km to the right. One of the limitations and weaknesses of eDNA is the low amount of extracted DNA, which may not be enough for amplification and comprehen- sive species detection, as well as the presence of inhibitors such as humic acid, which might affect the results [14]. Therefore, although generally accurate, eDNA results might not always reflect all species present and other factors should be considered [51, 52].

On the other hand, it is another interesting founding of present study that freshwater species were detected from adjacent coastal area by eDNA where freshwater species actually cannot exist. It can be referred that we collected eDNA moved by waterflow and the pattern differed by river (Fig 2). It would be worth to investigate about how far eDNA from river can transport to ocean and its relation with flow system. This could give a clue for eDNA localization in future study. Effect of human activity Human activity exerts a large influence on the environment and biodiversity [5, 6]. Water quality is closely related to the biodiversity of aquatic animals [9]. Using nutrient concentra- tions (TN and TP) and revetment rate as indices of human activity and urbanization, we deter- mined the impact of humans on biodiversity. GLM results indicated that TN significantly affected biodiversity, whereas the revetment rate had no effect (Table 3). The Tama River, which had the lowest biodiversity (Fig 2A), had the highest values for TN, TP, and revetment rates (Fig 3). The degree of urbanization of the Tama and Miya Rivers can be inferred not only from the concentration of nutrients and revetment rate but also from satellite images (Fig 1C and 1D). Even though the shoreline of the Sendai River has been extensively modified for flood control so that its revetment rate is now as high as for the Miya River, the surrounding area of the Sendai River has remained untouched and the nutrient concentration remains low

(Figs 1 and 3). The Miya River showed relatively high biodiversity because of its location in the southern part of Japan along the Pacific coast, which is affected by the Kuroshio warm current. In comparison, even though it is located in the northern part of Japan, biodiversity was quite high in the Aka River (Fig 2A), which can be explained by the vastly pristine environment of the river (Fig 1B). This is an important result as it indicates that efforts to conserve the envi- ronment can also improve biodiversity. Both the Takatsu and Sendai Rivers showed high bio- diversity with low human effect and geographical location (Figs 1 to 3). Conclusion The present study demonstrates that eDNA is a convenient tool for monitoring the distribu- tion, migration, and diversity of fishes. By simply collecting 1 L of water, we successfully detected 182 species including endangered species, covering a wide range of areas in a short period. Even though our experimental design has limitation that it is just a case study with PLOS ONE Biodiversity of five estuaries PLOS ONE | https://doi.org/10.1371/journal.pone.0231127 October 6, 2020 11 / 15