Multi-trophic eDNA and morphological monitoring reveal scale-dependent ecological responses to environmental stressors in arid-region rivers.

Arid-region rivers are increasingly affected by water scarcity, hydrological fragmentation, agricultural runoff, and urban development, creating an urgent need for monitoring approaches capable of resolving biodiversity responses across spatial scales. Here, we compared environmental DNA (eDNA) metabarcoding with conventional morphological monitoring across 40 sites in a highly regulated arid river-wetland system. Four biological groups were examined: eukaryotic algae, zooplankton, benthic macroinvertebrates, and fish. We evaluated differences between the two approaches in taxonomic records, community composition, spatial concordance, and sensitivity to surrounding-landscape and site-scale environmental stressors. eDNA metabarcoding recorded more taxonomic units than morphological surveys for eukaryotic algae, zooplankton, and benthic macroinvertebrates under the respective sampling workflows. However, the degree of agreement between the two methods varied among biological groups. Benthic macroinvertebrate communities showed relatively high compositional concordance, whereas planktonic communities exhibited pronounced spatial decoupling. Environmental modelling further revealed scale-dependent differences in model-dependent predictive importance. Biodiversity patterns derived from eDNA showed stronger associations with surrounding land cover within 2-km buffers, whereas morphological observations showed clearer associations with site-scale water-quality conditions, particularly the influence of dissolved oxygen on fish assemblages. These findings demonstrate that eDNA metabarcoding and morphological monitoring provide complementary rather than interchangeable information on aquatic biodiversity. Integrating the two approaches can improve the description of responses across biological groups and reduce spatial blind spots in fragmented arid rivers. Integrating the complementary information provided by the two approaches can strengthen biodiversity assessment across spatial scales and support more robust monitoring and management of arid-region river ecosystems.