Time‐Varying Storage–Water Age Relationships in a Catchment With a Mediterranean Climate

Abstract

Recent studies on the relationships between catchment storage and water ages using Travel Time Distributions (TTDs), Residence Time Distributions (RTDs), and StorAge Selection (SAS) functions have led to the hypothesis that streamflow preferentially mobilizes younger water when catchment storage is high. This so‐called “Inverse Storage Effect” (ISE) needs further evaluation in more catchments with diverse climates and physiographical features. In this work, we assessed the validity of the ISE in WS10 (H. J. Andrews forest, Oregon, USA), a forested headwater catchment in a Mediterranean climate. A conceptual model of the catchment, developed based on experimental observations of water flow paths in WS10, was calibrated to streamflow and δ18O in streamflow. Based on the calibrated model results, we determined RTDs, and streamflow TTDs and SAS functions by assuming that the soil reservoir and the groundwater reservoir act as well‐mixed systems. The streamflow SAS functions and travel time dynamics showed that the ISE generally applies in WS10. Yet, during transitions from dry summer periods to wet winter periods and vice versa, the marked seasonal climate caused rapid and strong storage variations in the catchment, which led to deviations from the ISE. The seasonality of streamflow travel times in WS10 is the result of the seasonal contributions of younger water from the hillslopes added to the rather constant groundwater contributions of older water. The streamflow SAS functions were able to capture the relative importance of contrasting flow paths in the soils and in the bedrock highlighted by previous studies in WS10.

The Demographics of Water: A Review of Water Ages in the Critical Zone

“…For conceptualizations of a hydrological system as a single (or several discrete) reservoir(s), the age master equation has been solved by specifying a "StorAge Selection (SAS) function" of all outflows, which can be used to represent different mixing assumptions inside the hydrologic compartment (Bolin & Rodhe, 1973;Rinaldo et al, 2015; Figure 3b). This approach can be applied to whole catchments (Benettin, Soulsby, et al, 2017;Harman, 2015;van der Velde et al, 2012), individual compartments (Benettin, Kirchner, et al, 2015;Rodriguez et al, 2018), and lakes (Smith et al, 2018). SAS functions are usually determined by assuming a functional form (e.g., a beta or gamma distribution) and calibrating the relevant parameters against observed tracer data, like stable isotopes (e.g., Benettin, Soulsby, et al, 2017) chloride (e.g., Benettin, Kirchner, et al, 2015), and recently also cosmogenic radioactive isotopes (Visser et al, 2019).…”

“…A number of recent studies calculating TTDs from catchment scale conceptual models have explicitly distinguished the unsaturated and saturated zones (e.g., Benettin, Bailey, et al, 2015;Benettin, Kirchner, et al, 2015;Birkel & Soulsby, 2016;Hrachowitz et al, 2013;Rodriguez et al, 2018). The conceptualization of the unsaturated zone and groundwater compartment was introduced to account for nonlinearity in the flow response, as the soil determines the temporally varying partitioning between storage, evaporation, transpiration, and recharge/drainage.…”

“…This recharge to groundwater can thus contain noticeable fractions of "old" water. The interactions of saturated and unsaturated zones are generally responsible for hysteresis in the stream water ages when plotted against catchment wetness (Benettin, Bailey, et al, 2017;Hrachowitz et al, 2013;Rodriguez et al, 2018;Yang et al, 2018).…”

Variability of Water Transit Time Distributions at the Strengbach Catchment (Vosges Mountains, France) Inferred Through Integrated Hydrological Modeling and Particle Tracking Algorithms

“…Up to now, studies about the temporal variability of TTDs and RTDs have mainly relied on concentration measurements of conservative geochemical or isotopic tracers [14], the underlying idea being to employ conceptual models or convolution kernels to determine the shape of the distributions that properly relate input and output geochemical signatures (e.g., [18,22,26]). Studies of distributions based on catchment flow and transport modeling are also available, but these have only received consideration as valuable alternatives to tracer-based studies [27][28][29][30][31][32][33][34]. The hydrological models used in these studies range from simply-conceived rainfall-runoff relationships (e.g., [30]) to more complex, fully distributed models (e.g., [31][32][33][34]).…”

“…Studies of distributions based on catchment flow and transport modeling are also available, but these have only received consideration as valuable alternatives to tracer-based studies [27][28][29][30][31][32][33][34]. The hydrological models used in these studies range from simply-conceived rainfall-runoff relationships (e.g., [30]) to more complex, fully distributed models (e.g., [31][32][33][34]). Transport processes are described using either conceptual approaches or explicit calculations such as the Eulerian resolution of an advection-dispersion equation or Lagrangian particle-tracking simulations.…”