Ecosystem Health and Sustainability · Published 2026-01-01 · DOI 10.34133/ehs.0551
Establishing ecological security patterns (ESPs) is critical for balancing regional economic development and ecosystem health. Current ESP frameworks predominantly focus on terrestrial ecosystems, neglecting aquatic ecological data integration. This study developed an innovative ESP approach by combining fish environmental DNA (eDNA) metabarcoding with Sentinel-2 remote sensing imagery in the Dongjiang River catchment. Two models were compared: Model 1 identified terrestrial ecological sources solely from InVEST outputs, while model 2 integrated riverine fish diversity (from eDNA) with 15 Sentinel-2 vegetation indices. A random forest classifier mapped continuous fish distributions, and principal-component-analysis-weighted environmental/anthropogenic variables generated a composite landscape-resistance surface. Circuit theory (via Linkage Mapper) extracted minimum-cost corridors between sources. Model 2 identified 109 habitat patches and 263 ecological corridors—markedly denser and more interconnected than model 1 (63 patches and 143 corridors). The integrated network demonstrated higher overall connectivity and maintained superior robustness under simulated habitat loss scenarios. These characteristics allow ecological processes and species dispersal to persist, enhancing the ecological resilience and functional stability of the network even under habitat fragmentation. Field eDNA surveys coupled with satellite imagery enabled precise identification of riparian ecological sources, while node/edge weighting corrected misjudgments of key source patches. Future work should optimize ESP models by incorporating hydrological and socio-economic variables to design comprehensive catchment-scale ESPs.
Abstract from DOAJ. Public domain (CC0 1.0).
Read the article at the publisher →