The analysis of hydrological and sediment dynamics during flood events offers a new perspective on material transport within basins and the management of rivers. The literature has highlighted the hysteretic nature of material transport at the event scale, which arises from temporal differences in the transport of materials (flow and sediment). Current hysteresis research methods fall short in effectively assessing the complex events. This study leverages suspended-sediment and discharge data from the upper Yangtze River to show proof-of-concept for hysteresis studies using Dynamic Time Warping (DTW) algorithm, a time series analysis tool. To evaluate the efficacy of the method, extensive assessments were undertaken using a combination of synthetic data and basin data with over 700 flood events. The proposed index effectively captures and quantifies hysteresis relationships associated with flood events, making it a versatile tool for hysteresis studies. We also similarly tested the robustness of the method and the application to complex events, and the results were satisfactory. Furthermore, we explored the hysteretic features of flood events in the studied basin. The results revealed that a proportion ranging from 50% to 60% flood events exhibit sediment peak leading pattern, indicating that in the upper Yangtze River, sediment peaks tend to occur in a leading pattern. The operation of reservoirs profoundly impacts the dynamics of sediment transport, resulting in an increased proportion (from 27% to 54%) of sediment peaks lagging behind discharge peaks. These study findings can be used to improve hysteresis study methods and contribute to the understanding of hydrological and sediment transport characteristics.

中文翻译：

---

一种分析暴雨事件排放-浓度滞后的新方法——以长江上游为例


洪水事件期间的水文和沉积物动力学分析为流域内的物质运输和河流管理提供了新的视角。文献强调了事件尺度上物质传递的滞后性质，这是由物质传递（流动和沉积物）的时间差异引起的。目前的磁滞研究方法在有效评估复杂事件方面存在不足。本研究利用长江上游的悬浮沉积物和排放数据，展示了使用时间序列分析工具动态时间扭曲 （DTW） 算法进行滞后研究的概念验证。为了评估该方法的有效性，结合使用合成数据和流域数据对 700 多起洪水事件进行了广泛的评估。拟议的索引有效地捕获和量化了与洪水事件相关的磁滞关系，使其成为磁滞研究的多功能工具。我们还同样测试了该方法的稳健性和对复杂事件的应用，结果令人满意。此外，我们探讨了所研究盆地洪水事件的滞后特征。结果表明，50%—60%的洪水事件表现出沉积物峰值领先型，表明长江上游沉积物峰值往往呈领先型。储层的运行对沉积物输送的动态产生了深远影响，导致沉积物峰值滞后于排放峰值的比例增加（从 27% 增加到 54%）。这些研究结果可用于改进磁滞研究方法，并有助于理解水文和沉积物输送特性。