Baseflow separation in a small watershed in New Brunswick, Canada, using a recursive digital filter calibrated with the conductivity mass balance method

Abstract

Baseflow separation is important for obtaining critical parameters for hydrological models. As measuring the baseflow component directly is difficult, various analytical and empirical baseflow separation methods have been developed and tested. The recursive digital filter (RDF) method is commonly used for baseflow separation due to its simplicity and low data requirement. However, parameters used in the RDF method are often determined arbitrarily, resulting in high uncertainty of the estimated baseflow rate. A more accurate method is the conductivity mass balance (CMB) method, which is established based on the differences in physical processes between baseflow and surface runoff. In this research, the output of the CMB method was used to calibrate the parameters of an RDF model, and the calibrated RDF model was used to estimate monthly, seasonal and annual baseflow rate and baseflow index for the past 19 years using streamflow discharge records. The characteristics of the baseflow hydrographs were found to be consistent with the hydrological and hydrogeological conditions of the research area. Research results indicated that the accuracy of the RDF model has been greatly enhanced after being calibrated with the CMB method so that the RDF model can provide more reliable baseflow separation results for a long‐term study. Copyright © 2012 John Wiley & Sons, Ltd.

Quantifying watershed‐scale groundwater loading and in‐stream fate of nitrate using high‐frequency water quality data

“…Tracer mass balance methods [Pinder and Jones, 1969], which rely on streamflow data and concentrations of chemical constituents in the stream, are also commonly used to separate hydrographs, and are often considered to be more objective than graphical hydrograph separation approaches [Stewart et al, 2007;Zhang et al, 2013;Miller et al, 2015]. We applied a graphical hydrograph separation approach to estimate base flow and runoff because the conductivity of stream water in the study streams is likely altered by application of road salts during high flow conditions in the winter months [Sanford et al, 2011], thereby limiting the use of a tracer mass balance approach to estimate base flow and runoff at these sites.…”

Promising new baseflow separation and recession analysis methods applied to streamflow at Glendhu Catchment, New Zealand

“…Empirical methods based on the hydrograph are the most widely used (Zhang et al, 2013), because of the availability of such data. The methods include (1) recession analysis (Linsley et al, 1975), (2) graphical methods, filtering streamflow data by various methods (e.g.…”

“…The methods include (1) recession analysis (Linsley et al, 1975), (2) graphical methods, filtering streamflow data by various methods (e.g. finding minima within predefined intervals and connecting them, Sloto and Crouse, 1996), (3) low-pass filtering of the hydrograph (Eckhardt, 2005;Zhang et al, 2013), and (4) using groundwater levels to calculate baseflow contributions based on previously determined relationships between groundwater levels and streamflows (Holko et al, 2002). One widely used empirical method for small catchments was proposed by Hewlett and Hibbert (1967) who argued that "since an arbitrary separation must be made in any case, why not base the classification on a single arbitrary decision, such as a fixed, universal method for separating hydrographs on all small watersheds?"…”

“…This has been a general feature in tracer studies and includes all of the components tested whether quickflow or baseflow (e.g. Hooper and Shoemaker, 1986;Bonell et al, 1990;Buttle, 1994;Gonzales et al, 2009;Zhang et al, 2013). In the case of groundwater, the rapid response is believed to be partially due to rapid propagation of rainfall effects downwards (by pressure waves or celerity) causing rapid water table rise and displacement of stored water near the stream (e.g.…”

“…0.8 for perennial streams in catchments with permeable bedrock). Eckhardt himself and others have pointed out that these BFI max values should be regarded as first approximations, and more refined values can be determined using tracers (Eckhardt, 2008;Gonzales et al, 2009;Zhang et al, 2013), by a backwards filtering operation (Collischonn and Fan, 2013) or by the relationship of two characteristic values from flow duration curves (i.e. Q 90 /Q 50 , Smakhtin, 2001;Collischonn and Fan, 2013).…”