Abstract: Four filtering method of Lyne-Hollick(L1), Chapman(L2), Chapman-Maxwell(L3), Boughton-Chapman(L4)were used for baseflow separation based on the data from Xilinhot hydrological station at Inner Mongolia, China. The paper explored the influence of different filter frequency and filter parameters on baseflow separation results and selected the optimal parameters to provide some experience for baseflow separation. The results showed that Boughton-Chapman filtering method with two filtering parameters had the best simulation effect, and the optimal segmentation parameter of Boughton-Chapman was n=4, f=0.95, Nash-Sutcliffe efficiency coefficient was 0.975, the average relative error was -0.3%, but it needed more separation times than that of the other methods. In the single parameter filtering methods, the Lyne-Hollick method had better simulation effect for baseflow separation, and the optimal segmentation parameter of Lyne-Hollick was n=3, f=0.8, Nash-Sutcliffe efficiency coefficient was 0.974, the average relative error was 3.95%. The results of Chapman and Chapman-Maxwell methods were relatively lower, and the average relative errors were greater than 10%.

Key words: Baseflow separation, Lyne-Hollick, Chapman, Chapman-Maxwell, Boughton-Chapman