BFlow and Eckhardt methods perform best among non-tracer-based baseflow separation methods
Qiaoqiao Wang, Qiang Li, Yanli Qin, Xuanjiao Chai, Hongli Chen, Yaping Wang, Mingfang Zhang, Dengfeng Liu
Abstract
Baseflow plays a crucial role in regulating streamflow, particularly during drought periods. However, selecting appropriate baseflow separation methods remains challenging due to varying parameter settings and their applicability to diverse hydrogeological conditions. Previous studies compared baseflow using non-tracer methods rather than tracer methods as a reference, often leading to biased estimates. Additionally, these studies relied on default or suggested recession constants provided by method developers without calibrating for watershed specific. To address these limitations, this study employs the conductivity mass balance (CMB) method to comprehensively evaluate non-tracer methods’ accuracy across diverse watershed and climatic conditions. We assessed four graphical and five digital filter methods (DFMs) using data from 131 watersheds in the Colorado and Mississippi River basins with available conductivity data. Our findings revealed significant discrepancies between graphical methods and the CMB method, highlighting their limited reliability for baseflow separation. In contrast, DFMs, particularly the BFlow and Eckhardt methods, demonstrated superior accuracy across various timescales and watershed characteristics, including elevation, vegetation cover, and hydrological regimes. Moreover, the incorporation of recession constants derived from recession analysis significantly improved the DFMs performance compared to suggested values. As such, we recommend the BFlow and Eckhardt methods for future baseflow-related studies. Overall, our study provided valuable guidance for selecting accurate baseflow separation methods and ultimately contributed to an understanding of baseflow dynamics and its role in watershed hydrology.