Stage‐Dependent Shifts in Controls on Annual Erosion‐Deposition Change Under Long‐Term Sediment‐Supply Decline in the Minjiang Estuary, Southeastern China

Under the combined influence of climate change and human activities, erosion–deposition patterns in many estuaries worldwide are undergoing substantial changes. However, under multifactor coupling, the drivers of estuarine erosion–deposition remain difficult to quantify; the dominant controls are not easily identified, and the underlying mechanisms are still insufficiently understood. Taking the Minjiang Estuary in southeastern China as an example, this study reconstructed subaqueous topographic surfaces from multiyear nautical charts, derived annualized DEM‐differencing erosion–deposition rates for 1992–1998, 1998–2007 and 2007–2019, and developed a stage‐specific interpretable machine‐learning framework to quantify the relative explanatory strength of different factors and examine their nonlinear driving mechanisms. The results show a clear stage‐specific reorganization of erosion–deposition controls in the Minjiang Estuary. Water depth remained highly important across all three stages, with normalized importance values of 24.0%, 26.1% and 20.5%, respectively, indicating a persistent constraint of subaqueous morphology on channel–shoal erosion–deposition patterns. Mean suspended sediment concentration was the dominant predictor during 1992–1998, with an importance of 32.9%, but declined markedly during 1998–2007 and 2007–2019, with importance values of 6.8% and 3.6%, respectively. This suggests that sediment availability shifted from a direct local explanatory signal to a broader system‐level background constraint. During 1998–2007, the importance of the engineering‐related obstruction metric increased from 9.6% to 20.7%, indicating an enhanced role of engineering structures in regulating the redistribution of limited sediment. The extended model for 2007–2019 further shows that mud content (11.4%), tidal current speed (10.1%) and significant wave height (10.1%) became important and exhibited nonlinear and condition‐dependent responses. Overall, after the construction of the Shuikou Dam in 1993, the Minjiang Estuary progressively shifted towards a sediment‐limited system, in which water‐depth morphology, human engineering activities, bed‐material conditions and tidal–wave energy jointly regulated erosion–deposition evolution. This study provides a transferable interpretable machine‐learning framework for diagnosing the time‐varying geomorphic controls of estuaries affected by the combined impacts of climate change and human activities.