Two techniques of tracking are compared for delineating the marine area impacted by a river plume. The first one uses a simple conservative tracer released in the river during a several year long period: the probability density function of the tracer concentration in each mesh of the model is used to map 3 characteristic descriptors of the freshwater influence (10%, 50% and 90% percentiles). The second one uses a signature of the river attached to a state variable of a biogeochemical model (e.g. the Dissolved Inorganic Nitrogen components), which can be followed in space and time in the various nitrogenous state variables. Departure between the two methods increases when tracking of nitrogen is considered in biogeochemical forms more and more distant from the original D.I.N one (e.g. zooplankton or benthic detrital N).