Unravelling the spatiotemporal variation in the water levels of Poyang Lake with the variational mode decomposition model

Min Gan,Xijun Lai,Yan Guo,Zhao Lu,Yongping Chen,Shunqi Pan,Haidong Pan,Ao Chu

DOI: https://doi.org/10.1002/hyp.15239

IF: 3.784

2024-07-16

Hydrological Processes

Abstract:The water level variations of Poyang Lake on diverse time scales, induced by human and natural factors, were identified using the variational mode decomposition model. The various effects of sand mining on the Poyang Lake water levels were detected. The Poyang Lake level's annual change is weakened when longer to the Yangtze River. The 3.6‐year change in the Poyang Lake levels is majorly caused by its tributary. Poyang Lake is a dynamic floodplain lake system that exhibits complex water level fluctuations and experiences significant regime changes over space and time, which remains to be further explored. This study used the variational mode decomposition (VMD) model to decompose the Poyang Lake's water levels from 1960 to 2022 at four key stations into six intrinsic mode functions (IMFs), namely IMF1–IMF6, representing variations on different time scales. The results present significant spatiotemporal heterogeneity. The multi‐year variation (IMF1) accounts for 5.6%–12.4% of the total variation and displays a northward decreasing trend, reflecting the lake's river‐like characteristics. The spectrum of IFM1 also reveals a significant 3.6‐year fluctuation mainly attributed to the tributary inflow, especially the Ganjiang River. The IMF1 differences between stations show abrupt decreases since the 2000s, indicating the impact of concentrated sand mining activities on the northern and central regions. The annual variation (IMF2) is the most prominent, contributing 76.1%–88.4% of the total variation, and shows a southward attenuation trend, likely due to the weakening influence of the Yangtze River flow. The intra‐annual scale (IMF3–IMF6) represents 6.0%–11.5% of the total variation and exhibits less spatial difference compared to the multi‐year and annual variations. The VMD model effectively separates the water level signals into different frequency bands, providing insights into the complex interactions between the lake, tributaries, and Yangtze River, as well as the impacts of human activities like sand mining, enhancing understanding of floodplain lake dynamics. The results also imply the importance of coping with the water level decline of Poyang Lake.

water resources