Decoding Marine Community Interactions and Stability through Environmental DNA Metabarcoding

Marine microbial communities are fundamental drivers of ocean biogeochemistry, mediating nutrient cycling, energy flow, and ecosystem functioning. Since their dynamics regulate the persistence and response of ecosystem processes to environmental changes, evaluating their stability is essential. Although the importance of marine microbial communities is widely acknowledged, most studies have focused on diversity patterns or compositional shifts rather than directly assessing stability and its underlying mechanisms. This gap is increasingly concerning in light of accelerating temperature variability, altered nutrient regimes, and shifting disturbance patterns. In this thesis, we integrate environmental DNA (eDNA) metabarcoding with ecological network analysis and nonlinear time-series modeling to investigate how environmental heterogeneity, interaction structure, and temperature-driven processes shape marine community stability across spatial and temporal scales. First, cross-domain networks reconstructed from eDNA metabarcoding using multiple genetic markers—capturing the broader trophic context influencing microbial dynamics—revealed that regional stability differences were better explained by network configuration and keystone taxa than by diversity alone. Communities with higher connectance and link density exhibited greater robustness. Second, long-term weekly eDNA data showed that free-living (FL) and particle-attached (PA) prokaryotic communities differ in seasonal network organization and stability, with biotic factors exerting a relatively stronger influence on PA stability and abiotic factors exerting a relatively stronger influence on FL stability. Third, empirical dynamic modeling of decade-long eDNA time-series revealed that temperature causally influences microbial abundance dynamics, with temperature-causal ASVs distributed across diverse prokaryotic lineages. Strong temperature-driven interactions were linked to increased abundance variability and reduced temporal stability, indicating taxon-specific differences in the capacity to buffer temperature-driven variability. Collectively, this thesis demonstrates how environmental variability and time-varying species interactions jointly regulate microbial stability by integrating eDNA-based community profiling, network-structural analysis, and causal time-series modeling. These findings enhance our predictive understanding of how marine ecosystems may reorganize in response to ongoing climate change.