3.1 Determination of a baseline period and trend analysis
Meteorological changes are an important driving factor of basin runoff changes. There are many methods to explore a trend-change in basin hydrological time series data but each method is based on a different principle. Due to uncertainty, a single method may cause unreliable results. The rationality of the results can be improved by analyzing the hydrological series of the basin in multiple ways. Lyu (2012) studied the hydrometeorological change trend of the Chabagou River Basin in China through Mann–Kendall (M–K) test and Spearman (Sp) method. This study uses the M-K test and Sp method to test the time series trend of hydrometeorology in the Qinhe River Basin.
Long-series hydrometeorological data are usually divided into multiple stages for comparative study. The division of stages is generally accomplished via two methods: 1) the ”manual division method” based on special years or special events (Miao et al., 2011) and 2) the ”breakpoint division method” based on different statistical methods (Tabari et al., 2014). The manual division method is simple but lacks specific division standards. The breakpoint division method is based on statistical methods to identify mutation years to divide different periods, which is more scientific. The traditional techniques for evaluating mutation points include the M-K test (Li et al., 2007), ordered clustering method (O-C) (Mohamed, 2017), and cumulative anomaly method (C-A) (Li and Song, 2020). In this study, the baseline period and change period were comprehensively divided using the above three methods.