| Abstract: | The movement and persistence of atrazine and metribuzin, in a sandy loam soil following application in spring, was simulated using two models. The first model, based on the physical laws describing water and solute movement and using measured values of soil hydraulic properties, underestimated herbicide mobility in the soil and predicted too rapid drying of the deeper soil layers. The accuracy of the simulations was improved by empirically reducing the measured hydraulic conductivities by a factor of 4. This probably reflects the difficulties of obtaining reliable measurements of soil hydraulic properties. A second and simpler model, which simulated water and herbicide movement using mobile and immobile water categories, accurately predicted soil water contents. It tended to underestimate herbicide movement at short times after application, and to overestimate movement later in the experiments. A comparison of different methods of simulating herbicide degradation showed that prediction of degradation rates in the field from laboratory data can be unsatisfactory with some compounds. |
