根据澄碧河流域平塘站 1963- 2017 年日径流资料, 采用最小滑动法、时间步长法和数字滤波法 3 类方法进行 基流分割研究。结果表明: ( 1) 基流总量方面, 数字滤波法对基流分割的结果优于最小滑动法和时间步长法, 其中, Boughto n2Chapman 滤波法分割的效果最好, Nash2Sutcliffe 效率系数为 01 930, 平均相对误差为 21 406% ; ( 2) 统计特 征方面, 3 类方法对不同水平年均有一定辨识能力, 且年代际变化基本一致, 呈波动趋势, 时间步长法偏差程度最 小, 最小滑动法和数字滤波法偏差程度大, 整体比较平稳; ( 3) 基流过程线方面, 3 类方法在汛期的表现差异较大, 非 汛期比较接近, 数字滤波法分割出的结果较平滑缓慢, 而最小滑动法和时间步长法分割出的结果拐点较多, 数字滤 波法中的 Boughto n2Chapman 滤波法比较符合基流迟滞特征。综上, Bo ug hton2Chapman 滤波法在澄碧河基流分割 中较为适宜。

This study applied the thr ee methods such as minimum smoo thing metho d, hy drog raph separatio n pro gr am method and dig ita l filter method to calculate baseflow using daily o bser ved runoff data from 1963 to 2017 at Pingt ang Station in Chengbi River Basin. T he results indicated that the dig ital filter metho d exhibited better perfo rmance than the minimum smo othing meth2 o d and hy dr og raph separat ion pr og ram method. The value of the err or ev aluatio n index es including t he Nash2Sutcliffe efficiency co efficient was 01 930 and the av erage relativ e err or w as 21 406% , respect ively. The results also demo nstr ated that the Boug hto n2 Chapman filter metho d had the hig hest simulatio n accuracy. In respect o f statistical cha racteristics, three kinds of metho ds had certain identificatio n ability in different year levels. T heir separatio n results appea red the same and stable fluctuat ion tr end at interdecadal scale. The dev iatio n deg ree o f the hy dr og raph separ atio n pr og ram metho d was smaller than that o f the minimum smo othing metho d and dig ital filter method. In terms o f discharg e hydrog ra ph, the perfo rmance o f the three methods was quit e discr epant in flood seaso n but relatively clo se in non2flood seaso n. The simulation results of the dig ital filter methods wer e smo other com par ed with t he o ther tw o ty pes of met ho ds. Part icularly, Bo ug hton2Chapman filter method in the dig ita l filter met ho d w as co nsistent w ith the hy steresis char acteristics of baseflow . In conclusion, t he Boug hto n2Chapman f ilter method w as more suitable in the division of Cheng bi Riv er baseflow .

国家自然科学基金( 51569003; 51609041 ) ; 广西自然科学基金( 2017GXNSFAA198361) ; 广西研究生教育创新计划资助项目 (YCBZ2019022, YCBZ2018023)