Ms Jia Lee1, Ms Youngeun Yang1, Dr. Yujin Kang1, Prof. Youngkeun Song1
1Seoul National University, , South Korea
Biography:
Jia Lee is a PhD candidate in the Interdisciplinary Program in Landscape Architecture at Seoul National University, Republic of Korea. Her research focuses on environmental DNA, biodiversity assessment, and conservation ecology, particularly the monitoring of amphibians with limited dispersal capacity in freshwater ecosystems. She earned her master’s degree in landscape architecture from Seoul National University, where she used environmental DNA to investigate fine-scale spatial variation in the distribution of the clawed salamander Onychodactylus sillanus.
Her current research applies eDNA metabarcoding, species-specific molecular assays, drone-based remote sensing, and spatial environmental data to biodiversity monitoring and ecological restoration planning. Her recent studies have addressed amphibian assemblages in abandoned paddy wetlands, fine-scale detection patterns of O. sillanus, and variation in eDNA detection across aquatic habitat types. She is also a co-inventor of a primer set developed for the molecular detection of Onychodactylus species.
Her broader research goal is to link molecular biodiversity data with landscape-scale environmental information to support evidence-based conservation, restoration monitoring, and adaptive management of freshwater ecosystems.
Abstract:
Environmental DNA (eDNA) surveys should capture species assemblages across contrasting environmental conditions. In small wetlands, fine-scale heterogeneity can be difficult to incorporate into sampling designs. We tested whether selecting sites to cover the range of environmental conditions identified from drone data could capture more of the amphibian species inventory than selecting the same number of sites at random. We used amphibian eDNA metabarcoding data from 85 water samples collected during five surveys at 17 sites in Pyeongdume Wetland, Mudeungsan National Park, Korea, in 2025. Environmental predictors included stream-network density and drone-derived measures of vegetation structure, vegetation vigor, and moisture processed using DJI Terra and LiDAR360. Using the complete dataset as a reference, we retrospectively evaluated every possible subset containing one to 17 sites. For each subset size, we identified combinations that covered the observed environmental range most broadly and compared their species coverage with that of all combinations containing the same number of sites. Gains in species coverage fell below two percentage points beyond seven sites. Seven-site combinations selected to cover the observed environmental range captured an average of 94.6% of the species detected across the full 17-site dataset, and 67.7% captured the complete inventory. In contrast, all possible seven-site combinations captured an average of 86.3%, and only 24.9% included every species. This advantage remained significant after accounting for multiple subset sizes (p = 0.049). These findings indicate that drone-derived environmental information can increase species inventory coverage without increasing the number of eDNA sampling sites in heterogeneous wetlands.