Michael K. Stewart

Modeling Base Flow Soil Water Residence Times From Deuterium Concentrations

Three approaches to determining mean soil water residence times in a steep headwater catchment were investigated. The deuterium concentrations of soil water collected from 11 suction cup samplers at the Maimai M8 catchment were determined weekly for 14 weeks and the results compared with those of rainfall in the same period. Deuterium variations in the suction samples were considerably delayed and diminished compared with the rainfall, indicating significant storage times and mixing with soil water. Soil matrix water at shallow levels (∼200 mm depth) in unsaturated soils was relatively responsive to fresh input, but deeper water and water near the stream subject to occasional water table rises showed much less variation. Steady state and non-steady state exponential models gave similar mean residence times, ranging from 12 to more than 100 days for different locations. Three groups of soil water response were defined, comprising shallow, medium and deep (near-stream) soil locations based on the mean residence times. The nonsteady models revealed considerable week-to-week and longer variations in mean residence time for shallow soil (SL4), but indicated that steady state models could adequately represent the system in the overall period investigated. In the third approach, model types and parameters that gave the best fits to the soil water deuterium concentrations were determined. Exponential and especially dispersion models were the most satisfactory. Weighting the input (rainfall δD) partially or fully with the amount of rainfall gave much worse fits than with the unweighted input, showing that much of the rainfall bypasses the soil matrix. The best fitting dispersion model (designated DM2) yielded the most accurate mean residence times: 13 days for shallow soil (SL4), 42 days for soil at 400 mm depth (SL5), both at midslope locations, and 63 days for soil at 800 mm depth near the stream (SL2). Capillary flow was important for the unsaturated shallow soil (SL4), while advection and hydrodynamic dispersion (mixing) were more dominant for the periodically saturated (SL5) and the generally saturated (SL2) soils.

Effect of bedrock permeability on stream base flow mean transit time scaling relationships: 2. Process study of storage and release

In Part 1 of this two-part series, Hale and McDonnell (2016) showed that bedrock permeability controlled base flow mean transit times (MTTs) and MTT scaling relations across two different catchment geologies in western Oregon. This paper presents a process-based investigation of storage and release in the more permeable catchments to explain the longer MTTs and (catchment) area-dependent scaling. Our field-based study includes hydrometric, MTT, and groundwater dating to better understand the role of subsurface catchment storage in setting base flow MTTs. We show that base flow MTTs were controlled by a mixture of water from discrete storage zones: (1) soil, (2) shallow hillslope bedrock, (3) deep hillslope bedrock, (4) surficial alluvial plain, and (5) suballuvial bedrock. We hypothesize that the relative contributions from each component change with catchment area. Our results indicate that the positive MTT-area scaling relationship observed in Part 1 is a result of older, longer flow path water from the suballuvial zone becoming a larger proportion of streamflow in a downstream direction (i.e., with increasing catchment area). Our work suggests that the subsurface permeability structure represents the most basic control on how subsurface water is stored and therefore is perhaps the best direct predictor of base flow MTT (i.e., better than previously derived morphometric-based predictors). Our discrete storage zone concept is a process explanation for the observed scaling behavior of Hale and McDonnell (2016), thereby linking patterns and processes at scales from 0.1 to 100 km.

River infiltration to a subtropical alluvial aquifer inferred using multiple environmental tracers

Chloride (Cl−), stable isotope ratios of water (δ18O and δ2H), sulfur hexafluoride (SF6), tritium (3H), carbon-14 (14C), noble gases (4He, Ne, and Ar), and hydrometry were used to characterize groundwater-surface water interactions, in particular infiltration rates, for the Lower Namoi River (New South Wales, Australia). The study period (four sampling campaigns between November 2009 and November 2011) represented the end of a decade-long drought followed by several high-flow events. The hydrometry showed that the river was generally losing to the alluvium, except when storm-derived floodwaves in the river channel generated bank recharge—discharge cycles. Using 3H/14C-derived estimates of groundwater mean residence time along the transect, infiltration rates ranged from 0.6 to 5 m yr−1. However, when using the peak transition age (a more realistic estimate of travel time in highly dispersive environments), the range in infiltration rate was larger (4–270 m yr−1). Both river water (highest δ2H, δ18O, SF6, 3H, and 14C) and an older groundwater source (lowest δ2H, δ18O, SF6, 3H, 14C, and highest 4He) were found in the riparian zone. This old groundwater end-member may represent leakage from an underlying confined aquifer (Great Artesian Basin). Environmental tracers may be used to estimate infiltration rates in this riparian environment but the presence of multiple sources of water and a high dispersion induced by frequent variations in the water table complicates their interpretation.