Modeling nitrate transport in watersheds using travel time distributions

Modeling nitrate transport in watersheds using travel time distributions Nitrogen-based fertilizers, used in intensive agriculture, and farming are the main causes of nitrate pollution of surface water. Nitrate excess is transported from lands to streams and lakes, leading to eutrophication, and thus to the depletion of water quality and ecosystems. At European level, increasing attention has recently been paid to this issue, as evidenced by the enactment of Nitrate Directive 91/676/EEC in 1991, with the aim of identifying the most endangered areas and implement the necessary countermeasures. Within this framework, new tools are needed, to understand and predict the fate of anthropogenic nitrate released on lands. This thesis was developed at the École Polytechnique Fédérale de Lausanne, as a part of the CH - International Mobility programme. It describes the set up of a model, implemented on Matlab, which computes non-conservative solutes concentration in the receiving water body draining a catchment, based on the hydrological fluxes (precipitation, evapotranspiration and discharge) and the inputs brought by human activities which act on the watershed's land. The model relies on a catchment-scale approach, which considers the watershed as whole, embedding its geology, hydraulic properties and climatic conditions in the integrated hydrological response at the outlet. The model is based on the Travel Time Distributions (TTDs) approach, which identifies each water volume by its age, i.e. the time spent within the catchment from its entrance as precipitation to its release as discharge. By means of StorAge Selection functions, it is possible to model the relative contribution of each water volume, characterized by a different water age, to the stream outflow. A mass-transfer equation which regulates solute exchange between soil and water is formulated. This links the age of a water parcel to its solute concentration and allows to calculate streamwater concentration. Two different formulations are developed, both for nitrate and geogenic solutes, found in water as a result of the weathering of minerals in the soil. Model performances are tested by means of data collected in a research catchment in France. Moreover, a calibration phase is conducted, to define the appropriate value of the parameters of the model which are site or solute-specific. Results of the simulations reported a remarkably good fitting with respect to geogenic solutes such as magnesium. As regards nitrate, the best results were obtained when taking into account only the periods characterized by a high flow rate. Indeed, nitrate streamwater concentration trend proved to be fully captured by the model upon the occurrence of high precipitation and discharge. Over the rest of the year, nitrate hydrological transport seemed to be regulated by different dynamics, probably due to an increasing importance of denitrification and to the occurrence of seep flows through preferential pathways. This study confirmed TTDs as a powerful tool to describe non-conservative solutes hydrological transport through watersheds. The developed model proved to have the potential to play an important role in the context of monitoring nitrate in land and water of catchments under cultivation. Future investigations could lead to the enlargement of the control volume and to the modeling of future scenarios, also with a view to evaluate the impact of climate changes.