The coefficient ae and the exponentbe (also known as the baseflow coefficients) can be derived using linear regression at log-log scale. Thebe value varies during the early to later recession state, where ae represents a function of the hydraulic characteristics controlled by factors including slope, topography, drainage system density (Zecharias & Brutsaert, 1988), aquifer thickness (Dewandel, Lachassagne, Bakalowicz, Weng, & Al-Malki, 2003), and hydraulic characteristics (Biswal & Marani, 2010), and others. The recession constant above can be also expressed in the form of the baseflow coefficients ae andbe (c = 1/[ae (2-be )],d= 2-be ). Related studies also made note of the correlation between coefficient ae and catchment drainage networks (Biswal & Marani, 2014; Shaw, 2016) and antecedent groundwater storage (Patnaik, Biswal, Kumar, & Sivakumar, 2015; Bart & Hope, 2014). Previous studies indicated recession parameters can represent the physical characteristics of the resulting aquifers and the structure of the river network and characterized catchment drainage behaviors (i.e. drainage characteristic timescale and flow recession rate (Brutsaert, 2008; Zhang, Chen, Hickel, & Shao, 2009; Troch et al., 2013; van Tol & Lorentz, 2018; Dwivedi et al., 2019). It also used to estimate hydraulic conductivity, drainable porosity, aquifer thickness, and length of stream channels (Oyarzún et al., 2014; Arumí, Troch, Maddock, Meixner, & Eastoe, 2016; Li, Zhang, Long, & Feng, 2017; Huang & Yeh, 2019). To the purpose of this study, we regarded the results as low-flow characteristics rather than defined the correlations with related physical characteristics.
The recession process may also be affected by land cover (i.e. vegetation coverage, vegetation types) and human activities (artificial pumping, agricultural return flow, etc.). Therefore, the water balance can be added an environmental impact during the recession period (Cheng et al., 2017):