“…Numerous time and frequency analysis methods are usually adopted to investigate the variations of flood event metrics at station scale, including temporal trend analysis methods (e.g., Mann-Kendall test, segmented regression model) (Shao and Campbell., 2002;Chen and Li, 2011a;Zhang et al, 2011b;Wang et al, 2015), change point tests (e.g., Pettit test, Mann-Kendall test, moving t-test) (Zheng et al, 2007;Li et al, 2008;She et al, 2017), periodicity analysis (e.g., wavelet analysis) (Anctil, 2010;Yang et al, 2018) and frequency analysis (e.g., copula analysis, entropy theory) (Zhang et al, 2011a;Singh, 2015). Most of existing studies mainly focus on flood magnitude metrics (e.g., flood volume, peak flood) and their frequencies at station scale (Milly et al, 2002a;Rogger et al, 2012;Xia et al, 2012;Ma et al, 2014;Requena et al, 2017;Merz et al, 2018;Sun et al, 2018). However, all of these findings are still difficult to reveal comprehensive variation characteristics at basin scale.…”