Answers are grounded in this paper and cite their sources.
Investigates influences on flood-to-drought transition in water levels, suggesting new management strategies.
Poyang Lake, the largest river‐connected freshwater lake in China, plays a vital role in maintaining regional ecological security and water resource balance; however, its flood‐to‐drought transition process has become increasingly pronounced, posing severe threats to wetland habitats and socioeconomic development. Existing studies fail to clarify the nonlinear interactive mechanisms between upstream inflows and the Yangtze River during the transition. This study used a Long Short‐Term Memory (LSTM) model and SHapley Additive exPlanations (SHAP) to explore the drivers of this transition from 1988 to 2018. Results showed that the LSTM model achieved high‐precision water level simulation, with Nash‐Sutcliffe Efficiency and Root Mean Square Error reaching 0.97 and 0.42 m for Xingzi Station (northern lake zone), 0.96 and 0.45 m for Tangyin Station (central lake zone), and 0.95 and 0.48 m for Kangshan Station (southern lake zone), respectively. SHAP analysis revealed that before the TGD, the Yangtze River dominated the flood‐to‐drought transition, and its influence was 385% that of the upstream rivers (Gan, Fu, Xin, Rao, and Xiu Rivers). After the TGD, the Yangtze's influence declined acceleratedly, while the upstream rivers' influence increased—even exceeding the Yangtze in central and southern zones. Furthermore, under the Yangtze River's modulation, the upstream inflow presented a clear critical discharge threshold for its marginal contribution to lake water level, with the effect shifting from no effective promotion of water level rise to significant lifting as discharge increased. This threshold decreased markedly after the TGD operation, indicating a significantly enhanced sensitivity of upstream inflow impacts on the lake's hydrological regime. This study quantifies altered interactive mechanisms of transition drivers, supporting the lake's water resource management amid flood and drought risks.
No takes yet. Share an insight, caveat, or question.
Jian et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: