（5）针对其他不同气候条件、纬度位置、地形地貌特征、人类活动过程地区，有待于进一步开展相关工作对本次研究结论加以验证、推广、完善。

      Accurate determination of base flow is of great significance in model calibration,water resources utilization, ecological water demand calculation, river sediment discharge calculation and river self-purification capacity calculation. However, it is difficult to observe the base flow directly, and there are some problems such as the lack of physical basis, difficulty in obtaining data, and significant distinctions in the baseflow separation results, so it is urgent to carry out comprehensive comparison and improvement research on the base flow separation methods. This work took the area located in the second Songhua River basin in China and the Mississippi River basin in the United States as examples to carry out a series of analyses. A variety of commonly used separation methods based on tracer and non-tracer were comprehensively compared, and their applicability and reliability in the above areas were analyzed. The following conclusions were drawn:

      (1) The ECK method has got the most reasonable result among various major hydrology data based baseflow separation methods, which can obtain the baseflow process that reflect the corresponding motion law perfectly, and can calculate the relatively stable baseflow index BFI. The two key parameters of ECK method, regression coefficient α and BFImax value, are suggested to be calculated by using the runoff regression curve method and the reverse filtering method, respectively.

      (2) The isotopic separation method can effectively identify the influence of the variation of water level difference value between surface water and groundwater and the change of runoff generation mechanisms on the baseflow. Furthermore, the isotopic separation results can be used to identify the ECK method parameters in different seasons, so that the optimal parameters in different seasons and different runoff generation mechanisms can be obtained, and the ECK method with these parameters can be applied to the calculation of subsequent runoff process in the basin, which can make the calculation more accurate.

     (3) The CMB method has a clear physical basis, and the monitoring data is easier to obtain than the isotopic method. The results indicated that the inverse correlation coefficient between discharge and conductivity can be used to quantitatively determine the applicability of the CMB method, while the CMB method is more applicable in anthropogenic activities. A minimum of six-month discharge and conductivity data was found to provide reliable parameters for the CMB method with acceptable errors, and it is recommended that the parameters SCRO and SCBF be determined by the 99th percentile and dynamic 99th percentile methods, respectively.

      (4) BRM method can effectively fit the shape of base flow process curve determined by tracer method, and the calculation result of this method is high precision. In particular, it can accurately calculate the river baseflow in a certain area, provided that parameters can be identified by using the CMB method according to the different seasons of each year.

      (5) Based on comprehensive comparison of various base flow separation methods, it was suggested that more than 6 months electrical conductivity monitoring process should be added on the basis of flow monitoring. The parameters of BRM method can be calibrated with the monitoring results, then more accurate baseflow separation results can be obtained by only use of flow data. When the conductivity monitoring conditions can not be met, recession coefficient α and BFImax, the two key parameters, can be obtained by using the dry season runoff regression curve method and reverse filtering method, and then the ECK separation method can be directly used to calculate the baseflow separation, which can also obtain a relatively accurate result.