Spatial-temporal correlation-based analysis of multi-source flood coincidence risks: A case study of the middle and lower Yangtze River basin

The middle and lower reaches of the Yangtze River basin, China. The overlapping of main and tributary river flood processes often leads to multi-source flood coincidence events, which involve complex spatiotemporal relationships and coincidence risks between different flood processes. This study analyzed flood spatial-temporal distribution patterns and multi-source flood components based on nearly 60 years of runoff data, and established a risk model for multi-source flood coincidences under spatiotemporal conditions using an optimal high-dimensional R-Vine Copula function, investigating the quantitative spatiotemporal relationships and changing patterns of risk in multi-source flood coincidences. It was found that multi-source flood coincidences are primarily composed of floods from tributaries of the Dongting and Poyang lakes, with the coincidence timing primarily concentrated in July. In the quantified spatial-temporal relationships of coincidences, the coincidence intensity index of floods from the Yangtze and Han Rivers is the highest, followed by that with the Dongting Lake, and finally with the Poyang Lake. Flood volume showing a stronger dependence structure compared to extreme flow sequences. Under the primary influence of the Three Gorges Dam, the risk of temporal coincidences with multi-source floods in the middle and lower Yangtze River has significantly decreased after the dam's operation. However, changes in the magnitude coincidence risk vary and are closely related to the flood recurrence interval, coincidence type, spatial-temporal distribution of rainfall and flood volume composition. • A flood coincidence risk model under spatiotemporal correlation is constructed. • Spatiotemporal quantitative relationship of flood coincidences is provided. • Three Gorges Dam operation can effectively reduce the temporal coincidence risk. • Three Gorges Dam operation doesn't reduce magnitude coincidence risk uniformly. • Rainfall pattern, flood volume, and interval flow affect magnitude coincidence risk.