金沙江下游梯级电站的建设抬升了河段水位，释放了河段的通航潜力。但电站日调节产生的非恒定流在下游河道传播中，使得坝下游水位波动频繁，传播特性复杂。为揭示水位波动驱动机制与传播特性，以乌东德坝下游约40 km河段为研究对象，通过原型观测等方法，系统研究河段水位变幅的时空分布规律及影响因素。结果表明：①坝下游水位波动呈现“非恒定流-库水位顶托”双向耦合驱动机制。②波动传播呈现“三阶段衰减”特性，近坝区（0~15 km）水位变幅衰减梯度最大；过渡区（15~35 km）因库水位顶托与流量波非线性叠加，衰减率波动性增强；远坝区（>35 km）完全受白鹤滩库水位控制，变幅趋于稳定。③白鹤滩库水位每下降10 m，非恒定流波峰传播距离平均增加2.3 km（R²=0.96），反映出下游反调节水库的削峰效应。④基于波动强度分区研究，提出“时空错峰”管理策略，建议通过航运区域动态管控、通航时段优化配置，将通航水域水位变幅控制在1 m/h的安全阈值内。

[Objective] This study aims to clarify the driving mechanisms and spatiotemporal propagation characteristics of water level fluctuations in the river section about 40 km downstream of the Wudongde hydropower station during its operation. The primary objectives are to： （1） quantify the dual impacts of unsteady flow from the hydropower station and the backwater effect from the downstream Baihetan Reservoir； （2） reveal the spatial attenuation patterns of water level variation； and （3） propose safety thresholds for water level variation in navigable waters. [Methods] Twelve high-frequency water level monitoring stations （SD1-SD12） were deployed along the river section 40 km downstream of the dam， and water level data were collected at 5-minute intervals. Four intensive monitoring activities were conducted， covering key operational periods including the impoundment period， drawdown period， flood season， and low-water level operation of the Baihetan Reservoir. The correlations among flow variation from the Wudongde hydropower station， water level of the Baihetan Reservoir， and downstream water level variation were quantified using statistical methods， including linear regression and correlation analysis. In addition， the spatial gradients and temporal attenuation of daily and hourly water level variation were calculated and analyzed. [Results] 1） The water level fluctuations downstream of the Wudongde Dam exhibited a bidirectional coupled driving mechanism， consisting of unsteady flow and reservoir backwater effects. In the near-dam reach （0-15 km）， the station’s flow was the primary driver of water level variation， with the water level response coefficient remaining stable at 0.12-0.15 m per 100 m³/s （R²>0.99）. Conversely， the far-dam reach （> 35 km） was dominated by the water level of the Baihetan Reservoir， exhibiting slow-varying， reservoir-controlled behavior with relatively stable water levels. 2） Downstream water level fluctuations exhibited a three-stage attenuation pattern. In the near-dam reach （0-15 km）， water level variation decreased rapidly， with a maximum daily attenuation rate of 0.41‰， and was strongly correlated with flow fluctuations （R²=0.85）. In the transition zone （15-35 km）， the interaction of nonlinear waves increased the variability of the attenuation rate （0.02‰-0.31‰ per day）， leading to a sharp weakening or disappearance of the correlation with flow. In the far-dam reach （>35 km）， the backwater effects of the Baihetan Reservoir stabilized water levels. 3） Water level regulation at the Baihetan Reservoir exerted a dampening effect on wave peak propagation. For every 10 m decrease in reservoir water level， the propagation distance of the flow peak increased by an average of 2.3 km （R²=0.96）. The nighttime flow peak at the Wudongde Dam occurred between 16：00 and 24：00， and flow levels exceeded daytime peaks by more than 40%. When the daily and hourly flow variations at Wudongde exceeded 5 000 m³/s and 1 500 m³/s， respectively， the resulting water level variation （3.2 m/d and 1.2 m/h） exceeded the shipping safety thresholds （3.0 m/d and 1.0 m/h）. [Conclusion] Downstream water level fluctuations are governed by a bidirectional coupled mechanism of unsteady flow and reservoir backwater effects， and exhibit a distinct three-stage attenuation pattern. This study identified the nonlinear superposition phenomenon in the transition zone （15-35 km） for the first time based on prototype observations， revealing that opposing wave phases （flow waves and backwater waves） generated complex peak interference effects. Furthermore， navigation safety thresholds are determined， and a “spatiotemporal peak-staggering” management strategy is proposed. This strategy involves dynamic zoning control， such as restricting navigation in the high-risk near-dam reach （0-15 km， SD1-SD4） during the peak period of 16：00-24：00.