Combining Hydrometric and Hydrochemical Data Sets for Investigating Runoff Generation Processes: Tautologies, Inconsistencies and Possible Explanations

Abstract

The rapid increase of catchment runoff after the onset of rainstorms and the associated short‐term changes of stream‐water solute concentration are still poorly understood. These factors are dependent on the flowpaths that the water follows to the stream, and the length of time from rainwater infiltration to discharge. To study these processes, different approaches are used. Hydrometric data on, for example, discharge, groundwater level, soil matrix potential and soil moisture data, have a high spatial and temporal resolution. In contrast, hydrochemical data for solutes, isotopes and temperature, are spatial and temporal integrals. This article provides an overview of the different approaches that are used to investigate stormflow runoff, and critically examines apparent and actual tautologies and inconsistencies in them. Some of the inconsistencies can be traced back to a failure of the basic assumptions, whereas others remain to challenge our understanding of the runoff generation process. Finally, suggestions for future avenues of research in this field are made.

Towards a simple dynamic process conceptualization in rainfall–runoff models using multi‐criteria calibration and tracers in temperate, upland catchments

“…This provided an additional diagnostic tool to explore how well the model was working (Kirchner, 2006). As noted by Lischeid (2008), the nature of the 'quick' and 'slow' stores controlling catchment response can be examined by empirical data such as tracers and field surveys of saturation area dynamics. However, spatial and temporal variability in soil and groundwater chemistry in the Girnock are marked, and it is not easy to capture this in a simple modelling approach .…”

“…Moreover, uncertainties relating to model structures (Beven, 1993), parameter identifiability (Son and Sivapalan, 2007), equifinality (Beven, 2006), accuracy of calibration data (Fenicia et al, 2007) and spatial heterogeneity (Bloeschl and Sivapalan, 1995) have demonstrated that runoff simulations alone can be a poor test of whether a model has adequately captured the important controlling processes (Lischeid, 2008). This has to be taken into account for many applications where key processes need to be incorporated, for example prediction of climate change or diffuse pollution (e.g.…”

Hydrological connectivity inferred from diatom transport through the riparian-stream system

“…We compare diatom results with traditional two-component hydrograph separation, and end-member mixing analysis (EMMA) using stream water chemistry and stable isotope data. We also present soil water content and groundwater level data within the HRS system to facilitate a somewhat holistic understanding of catchment runoff processes (as advocated by Bonell, 1998;Burns, 2002;Lischeid, 2008). Specifically, we addressed the following questions.…”

Multi-criteria parameter estimation for the Unified Land Model

“…Gupta et al, 1999;Crow et al, 2003;McCabe et al, 2005;Khu et al, 2008;Werth and Güntner, 2010;Milzow et al, 2011). Among the variables other than streamflow that have been used for LSM evaluation are evapotranspiration (Nandagiri, 2007), surface heat fluxes (Gupta et al, 1999;McCabe et al, 2005), hydrochemical and isotope tracers (Son and Sivapalan, 2007;Lischeid, 2008;Birkel et al, 2010), land surface temperature (Crow et al, 2003;McCabe et al, 2005), remotely sensed soil moisture (Brocca et al, 2010;Milzow et al, 2011), snow water equivalent (MacLean et al, 2010), terrestrial water storage (Werth and Güntner, 2010;Milzow et al, 2010), and water table level (Khu et al, 2008). The more frequent use of multivariate observations is attributable in part to their growing availability.…”