Experiments in hydrologic solute transport

This thesis deal with experimental, field and modeling studies on reactive solute transport within hydrologic control volumes. The substances used in open environments for plant protection and other purposes are likely to be flushed by rainfall and enter into the hydrologic system if they are not disposed of through a sewage network. This represents a significant hazard for receiving ecosystems. The concentrations observed in surface waters are often highly dynamic, because of the variability of the hydrologic drivers controlling solute transport. This makes surface water monitoring particulary laborious, thus calling for a better understanding, and description of hydrologic transport at relevant spatial and temporal scales. The features of hydrologic transport are investigated under controlled conditions in a large weighing lysimeter, where all input (rainfall) and output (evapotranspiration and bottom discharge) fluxes are closely monitored. Random rainfall and large evapotranspiration (ET) fluxes dictated by two willows induce highly variable hydrologic conditions in the system. Fluorobenzoic acids (FBAs) as well as isotopically labeled water were used to unequivocally mark selected rainfall events. Chemical analyses were performed on samples of discharge and soil water collected through sampling ports in order to estimate the breakthrough curves of each tracer. Besides revealing unreported reactive behavior of the FBAs, these results demonstrate non stationary tracer responses that not only result from the transient precipitation forcing, but also from the variability in the ET-induced water deficits and how the output fluxes sample the water and solutes in storage under the prevalent moisture conditions. The experimental data were further explored using a model based on travel time distributions (TTDs). Describing hydrologic transport using TTDs is becoming increasingly common in catchment hydrology, because it allows a collective measurement of many processes (climate forcing, internal mixing, flow pathways, etc.) into a single stochastic descriptor, therefore removing the need to estimate numerous physical parameters and circumventing the limited applicability of point-scale laws for structurally complex and heterogenous environments. This parsimonious model is able to reproduce the different tracer responses by keeping track of the age composition of the stored water and assessing the ways ET and discharge sample the available water ages in time. The results emphasize the effects of ET on solute transport, because it samples water of different ages (i.e. residence times) in the transport volume in regards to discharge and thereby strongly modifies the resulting TTDs. In this way it is possible to compare the respective roles of the hydrologic variability and reactivity attributes (namely plant uptake and microbial degradation) which affect the tracer responses and prevent a direct interpretation of the experimental breakthrough curves as