Modelling pesticide sorption in the surface and subsurface soils of an agricultural catchment.

Background: Sorption models that improve upon the koc concept are urgently needed for reliable spatial modelling of pesticide leaching. Sorption of glyphosate, bentazone and isoproturon was measured in surface and subsurface soils to test an 'extended' partitioning model that also accounts for inorganic sorbents and pH. Best-subset regression and Akaike information criteria were used to justify the inclusion of predictors and identify suitable models.

Results: The extended partitioning model improved upon the koc concept for all three compounds: inorganic sorbents dominated sorption in subsurface soils, and their effects were only masked by organic matter in surface soils with organic carbon contents larger than ca 2%. Interactions between organic and inorganic sorbents affected glyphosate sorption, but apparently not that of bentazone or isoproturon.

Conclusion: Information on clay, iron and aluminium oxides and soil pH, in addition to organic carbon, is needed for accurate prediction of pesticide leaching. The variables foc , fclay and pH are generally available, whereas measurements of oxides of Al and Fe are rarely reported. The authors therefore emphasise the need to measure and report contents of oxides of Al and Fe in soil survey databases, because small variations in their concentrations may contribute significantly to large variations in sorption, especially of ionisable pesticides.