2020 年 26 巻 p. 319-324
The applicability of the environmental DNA metabarcoding (MB) assay for fish surveys was investigated using case studies of the National Census on River Environments. The fish lists compiled using MB or traditional sampling (TS) were compared based on 188 fish taxa from 76 sites from 15 case studies. The detection rate by MB was high for migratory swimming fishes (92.9 %) and low for benthic marine/brackish fishes (59.2 %). The fish taxa were detected at a probability of 76.9 ± 17.1 % (mean ± SD) by the MB, and 74.2 ± 16.5 % by the TS. The agreement rate between the lists from both methods was 51.1 ± 16.4 %, and a positive correlation was obtained for the 67 of the 76 sites (p < 0.05, cor-2 test). The sites with no correlation had one of the following characteristics: (1) less than 10 taxa, (2) major gap in the number of taxa, (3) located in estuarine area, and (4) implementation timing was mismatched. The differences in the taxon lists among multiple sites on the same river correspond to the variation in environmental status.