Can eDNA Replace Trawl Surveys for Estuarine Species Distribution Modeling: Insights From Collichthys lucidus in the...

Species distribution data underpin species distribution models (SDMs), which are essential for identifying habitat preferences and informing conservation strategies. Environmental DNA (eDNA) has emerged as a powerful tool for aquatic biodiversity monitoring. However, its reliability in supporting SDMs-especially in dynamic estuarine systems-remains uncertain. To address this, this study evaluated the modeling performance of trawl-derived and eDNA-derived occurrence data for the benthic fish Collichthys lucidus, using paired surveys conducted in the Yangtze River Estuary in August and November 2021. Species distribution models were developed using both MaxEnt and ensemble modeling (EM) approaches. The results showed that although eDNA and trawl data produced similar performance metrics (AUC, Kappa, RMSE), eDNA-based models exhibited weaker spatial discrimination and inconsistent seasonal predictions, misidentifying offshore areas as unsuitable under certain conditions. In contrast, trawl-based models consistently identified suitable offshore habitats and highlighted salinity as the dominant driver, whereas eDNA models emphasized dissolved oxygen. These findings suggest that using eDNA as a sole input for spatially explicit habitat modeling in estuaries requires caution, particularly where precise location information is critical.