Abstract: Analysis on the runoff evolution patterns of water sources along the Pingtan and the Minjiang River Estuary Water Resources Allocation Project (the "One Gate Three Lines" Project) is of great significance for optimizing the spatially balanced allocation of water resources in Fujian Province.Based on long-term runoff data spanning six decades from five major water sources along the project route (Yongtai Station, Zhuqi Station, Sanxikou Reservoir, Dongliu Reservoir, and Sanshiliujiao Lake), we employed the Mann-Kendall rank correlation test, Pettitt change-point detection, wavelet analysis, concentration degree (CD), and concentration period (CP) to systematically characterize inter-annual and intra-annual runoff distribution patterns among these water sources.In addition, a Copula function was used to analyze the wetness-dryness encounter characteristics among water sources.The results showed that: ① in terms of inter-annual variations, the annual and flood-season runoff at Yongtai Station exhibited a significant decreasing trend and experienced a significant abrupt change in 2000;whereas the non-flood-season runoff at Zhuqi Station showed a significant increasing trend, with its abrupt change point identified in 1981;② wavelet analysis revealed dominant periodicities of 8~18 years for both annual and flood-season runoff across all water sources, while non-flood-season runoff exhibited primary cycles spanning 5~13 years; ③ regarding intra-annual distribution, runoff at all water sources was concentrated from May to September, with CD values ranging from 0.40 to 0.53, indicating a relatively uneven distribution.Notably, the runoff CD at Yongtai and Zhuqi Stations exhibited a declining trend, while the CP showed a potential trend toward delay; ④ regarding wetness-dryness encounter characteristics, significant hydrological compensation patterns were observed among different water sources, which provide favorable conditions for implementing multi-source joint operations, with synchronous occurrence probabilities ranging from 0.47 to 0.65 and asynchronous probabilities ranging from 0.35 to 0.53.These findings can provide a scientific foundation for formulating coordinated multi-source operation strategies in trans-basin water transfer projects.