Abstract. Streamflow response to precipitation is often investigated using isotope-based hydrograph separation, which quantifies the contribution of precipitation (i.e., event water Qe) or water from storage (i.e., pre-event water Qpe) to total discharge (Q) during storm events. In order to better understand streamflow generating mechanisms, hydrograph separation studies often seek to relate the event water fraction Qe / Q to storm characteristics or antecedent wetness conditions. However, these relationships may be obscured because the same factors that influence Qe also necessarily influence total discharge Q as well. Here we propose that the fractions of event water and pre-event water relative to precipitation instead of discharge (Qe / P and Qpe / P) provide useful alternative tool for studying catchment storm responses. These two quantities separate the well-known runoff coefficient (Q / P, i.e. the ratio between discharge and precipitation over the event time scale) into its contributions from event water and pre-event water. Whereas the runoff coefficient Q / P quantifies how strongly precipitation inputs affect streamflow, the fractions Qe / P and Qpe / P track the sources of this streamflow response. We use high-frequency measurements of stable water isotopes for 24 storm events at a steep headwater catchment (Erlenbach, central Switzerland) to compare the storm-to-storm variations in Qe / Q, Qe / P and Qpe / P. Our analysis explores how storm characteristics and antecedent wetness conditions affect the mobilization of event water and pre-event water at the catchment scale. Isotopic hydrograph separation shows that streamflow was typically dominated by pre-event water, although event water exceeded 50 % of discharge for several storms. No clear relationships were found linking either storm characteristics or antecedent wetness conditions with the volumes of event water or pre-event water (Qe, Qpe), or with event water as a fraction of discharge (Qe / Q), beyond the unsurprising correlation of larger storms with greater Qe and greater total Q. By contrast, event water as a fraction of precipitation (Qe / P) was strongly correlated with storm volume and intensity but not with antecedent wetness, implying that the volume of event water that is transmitted to streamflow increases more-than-proportionally with storm size under both wet and dry conditions. Conversely, pre-event water as a fraction of precipitation (Qpe / P) was strongly correlated with all measures of antecedent wetness but not with storm characteristics, implying that wet conditions primarily facilitate the mobilization of old (pre-event) water, rather than the fast transmission of new (event) water to streamflow, even at a catchment where runoff coefficients can be large. Thus, expressing event and pre-event water volumes as fractions of precipitation rather than discharge was more insightful for investigating the Erlenbach catchment's hydrological behaviour. If Qe / P and Qpe / P exhibit similar relationships with storm characteristics and antecedent wetness conditions in other catchments, we suggest that these patterns may potentially be useful as diagnostic fingerprints of catchment storm response.
A high-resolution global runoff estimate based on GIS and an empirical runoff coefficient
