Field water and solute fluxes : Optimizing sampling frequency and calculation procedures

Leaching of large quantities of fertilizers is an important source of groundwater pollution. This nonpoint source pollution must be estimated when designing improved agricultural systems. One indirect estimation method for solute fluxes is the coupling of solute concentrations with water fluxes obtained from a water budget. In this article, we investigate the possibility of using this method with solute concentration obtained from suction lysimeters installed at four different depths and water fluxes calculated from time domain reflectrometry and meteorological data. The precision of this method was assessed using three pan lysimeters installed with limited soil disturbance. The solutes considered were nitrate and bromide. The precision on the water flux estimated by the water budget increased with the time interval used for calculation. A relative mean absolute error (MAE) of 10% on drainage amounts was obtained for sampling periods longer than 20 days. Better agreement of measured fluxes was obtained with nitrate compared to bromide, probably because bromide was not uniformly distributed vertically within the pan lysimeters. Better estimated solute fluxes were obtained by coupling estimated drainage to measured bromide concentrations in the suction lysimeters of the lowermost soil layer. When using nitrate concentrations from the lowermost layer, the precision of nitrate fluxes was increased for longer periods between sampling; precision increased only marginally after 12 days with MAE values of 98%, 40%, and 90%, respectively for pan lysimeters A, B, and C. These MAE values are larger than those for drainage where they never exceeded 25% at 12 days. The larger errors on nitrate fluxes, compared to water fluxes, are due to errors on nitrate concentration that increase with time between consecutive samplings. Such an error on the nitrate fluxes can be decreased to about 15% if frequent concentration measurements (e.g., 1, 2, or 3 days) were taken, and nitrate masses were cumulated over periods of at least 20 days.