Modeling Nitrate export with GR4J and StorAge Selection functions

Models based on StorAge Selection (SAS) functions are emerging as promising tools for quantifying catchment-scale solute export through the understanding of factors controlling transit time distribution (TTD). SAS functions describe the sampling of different ages in catchment storage that produces river discharge. This sampling varies over time, for example in such a way that the young water fraction is generally higher during high flows and lower during low flows. Literature thus proposed different parametrization of SAS functions in order to make them dynamics according, for instance, to soil moisture or antecedent inputs.  

In the first step of this work, we investigated how the time variance of the SAS functions could be related to the different fluxes and model states of the hydrological model GR4J. Chloride concentrations are used as a conservative tracer to evaluate different modelling hypotheses. In a second step, we investigated how TTD could be used to understand the dynamics of a non-conservative solute such as nitrate. In addition to the solute transport, we described the relationship between water age and denitrification at catchment scale. We assumed that nitrate removal is higher when young water fraction is lower, and we investigated the seasonal dynamics of this assumption. These hypotheses are tested by exploring the detailed water quality record from the Orgeval catchment (ORACLE observatory, France) where it is possible to assess various scales of nested catchments (up to 104 km²).