Comparison of the detection of 3 endangered frog species by eDNA and acoustic surveys across 3 seasons

Abstract

For conservation of endangered species, basic ecological information is indispensable. Environmental DNA (eDNA) is increasingly used to estimate the presence/absence of aquatic organisms such as frog species. However, the efficiency of detecting eDNA may depend on the ecological characteristics of the target frog species. Here, we adopted an eDNA approach for detecting 3 endangered Japanese species (Babina subaspera, Odorrana splendida, and O. amamiensis). We compared eDNA detection with a traditional acoustic survey method in 3 different seasons. Detection of species was not always consistent between the 2 methods, probably because acoustic surveys target only calling adult males but eDNA detects individuals of both sexes and at any life-history stage in water. Moreover, the differences in detection patterns between the 2 methods among species may be correlated with different characteristics of mating behavior (e.g., calling volume) and life history of tadpoles (e.g., migration ability). The optimal season for detection by eDNA method also differed between the 3 species, probably because of differences in breeding season and larval characteristics. We show that results obtained from eDNA detection methods and traditional field observation techniques, including acoustic surveys, are not always consistent with each other but rather depend on the ecological characteristics of the respective target species. It is therefore necessary to consider the ecological traits of frog species, such as breeding season and seasonal shifts in tadpole abundance, when applying eDNA methods for distribution surveys.