**1. Introduction**

Biological monitoring contributes to the understanding of complex ecosystem structures and functions in targeted systems [1]. Accordingly, it is crucial in detecting and assessing environmental changes in order to ensure proper managemen<sup>t</sup> and conservation of complex ecosystems [2]. Coastal environments are among the most complex ecosystems due to tidal activity, and typically retain high economic and environmental values in light of aquatic resources and biodiversity [3]. Coastal ecosystems are often severely a ffected by anthropogenic activities such as industrial fishing [4,5], marine transport and leisure activities [6], aquaculture and the aquarium trade [7,8], living seafood and lure fisheries [9,10], and non-indigenous species (NIS) which induce greater pressures on endemic ecosystems and often drive native species to extinction in the resulting habitats [11–13]. The nonlinearity and complexity of coastal ecosystems due to the aforementioned activities often causes di fficulties when analyzing spatial and temporal patterns of ecological traits.

Environmental DNA (eDNA) as described by Ogram et al. [14] who extracted microbial DNA from the sediment, has increasingly been used in recent years for biological monitoring purposes. Recently, a large number of papers have reported the use of eDNA monitoring in analyses of soil, water and even air [15]. Andersen et al. [16] examined the possibility of monitoring large mammals using eDNA in soil samples, and eDNA from water monitoring of fish [17–20] and amphibians [21,22] has been successful. Furthermore, Hawkins et al. [23] demonstrated that a complete taxonomic list of functional feeding group (FFG) criteria, based on high resolution of identification (genus or species level) based on DNA techniques, can determine the e ffects of watershed alterations on stream invertebrate assemblages in bulk eDNA samples. However, family level identification based on visual inspection did not reveal any di fferences of FFG composition between sites. Next-generation sequencing (NGS) technologies for eDNA monitoring have provided an alternative to overcome issues such as identification problems associated with classical monitoring in a species rich coastal environment [3,20].

The values of coastal ecosystems, such as primary production (i.e., sea grass and algae) and commercial fish yields, are intertwined with multiple environmental factors, including nutrient concentrations (carbon, nitrogen, and phosphorus), phytoplankton growth, zooplankton grazing effects, and benthic communities. The Gwangyang Bay coastal ecosystem is the most economically productive in Korean peninsula. Specifically, in the midst of three major cities (Gwangyang, Yeosu and Suncheon) in Jeonnam Province, with Gwangyang Bay, it yields 71% (1,297,815 tons) of the annual aquacultural resource output as of 2016 [24]. However, there is a large industrial area near the Bay, and the area is primarily involved in industrial activities such as oil refineries and steel production plants. Kim et al. [25] characterized the dissimilarity of water quality and sediment contamination, and identified the importance of nutrients supplied by rivers. Such findings are, however, still limited to representing general ecological features of the Gwangyang Bay coastal ecosystem.

The main objective of this study was to determine aquatic community taxonomic composition using eDNA based on an NGS approach for characterizing general ecological features in the Gwangyang Bay coastal ecosystem. We analyzed the community spatial distribution with regard to environmental parameters, and the habitat types (marine, freshwater and estuarine), feeding habits (filter feeder, carnivore, producer and symbiotic) and indigenous species rate (ISR) among the three di fferent zones referred to by Kim et al. [25]. Moreover, we discuss the e ffectiveness and sensitivity of our NGS approach on the Gwangyang Bay coastal ecosystem.
