Modelling metolachlor exports in subsurface drainage water from two structured soils under maize (eastern France)

The study investigated the processes involved in metolachlor transport in two artificially drained, structured soils in eastern France. Measured losses of bromide and metolachlor in drainage water were compared with results simulated by the mechanistic, stochastic AgriFlux model. Simulated drainage water volumes were generally similar to the measured volumes when the spatial variability of the soil water properties was taken into account. When such variability was disregarded, cumulative water volumes of the clay soil were over- or underestimated by more than 20%. Two types of adsorption were tested. For instantaneous, reversible adsorption, using the partition coefficient Koc, metolachlor losses were underestimated in the first drainage water volumes and overestimated for the total study period. The use of slow adsorption and desorption kinetics (ADK) produced an export pattern similar to the observed one. A sensitivity analysis indicated that the simulated results are very sensitive to the values of the ADK rates, especially for the silty loam soil. The effect of ADK on the attenuation of metolachlor exports was more significant than the effect of degradation (2.3 and 6.7 times higher for the clay and silty loam soils, respectively). For the same four-month period, the bromide and metolachlor losses (using ADK) in the clay soil were 2.1 and 1.3 times greater, respectively, if the macroporosity was set at 10% than if it was not simulated. Conversely, macroporosity did not significantly affect these losses in the silty loam. The main factors involved in the metolachlor transport in the studied soils were: (i) the macroporosity, especially in the clay soil because of the low hydraulic conductivity of the matrix and (ii) the sorption kinetics rates which varied according to the soil physico-chemical characteristics.