Long‐Term Water‐Level Dynamics and Regime Shifts in a Regulated Irrigation System: Evidence From the Bac Hung Hai Network, Vietnam

This study investigates long‐term water‐level dynamics in the Bac Hung Hai irrigation system, Red River Delta, Vietnam, based on continuous six‐hourly observations from 11 monitoring stations over 26 years (2000–2025). An integrated framework, incorporating seasonal trend decomposition based on Loess, Mann–Kendall and Seasonal Mann–Kendall trend tests, Sen's slope estimation, Bayesian change‐point detection and Empirical Orthogonal Function analysis, was employed to characterise variability, long‐term trends, regime shifts and spatial coherence. Results indicate a systemic decline in water levels, most pronounced at the upstream intake (−90 mm year −1 ), which is temporally consistent with Red River bed incision and upstream reservoir regulation, though direct causal attribution requires boundary‐condition modelling. Bayesian analysis identifies a system‐wide hydraulic regime shift around 2009–2010, coinciding with the commissioning of the Son La reservoir and marking an operational shift consistent with increased pump dependence. EOF analysis demonstrates that 83% of network variance is explained by two modes: a coherent basin‐scale mode (69%) and an upstream–downstream contrast mode (14%), which reflects the spatial tension between supply‐restricted intake and drainage‐intensive outlet zones. These findings demonstrate that the BHH hydraulic regime is nonstationary and operationally fragmented, with implications for the adaptive management of regulated lowland irrigation systems.