Seven pesticides, 2,4-D, atrazine, hexazinone, picloram, procymidone, terbuthylazine, and triclopyr, were applied to a Horotiu soil (allophanic loamy silt) located near Hamilton, New Zealand, along with 2 tracer compounds, bromide and deuterated water. Their movement and persistence was monitored for about 2�years using soil sampling down to a maximum of 1 m and 9 suction cups located between 0.2 and 2.5 m down the profile. There was rapid leaching of the tracers as well as hexazinone, picloram, and, to a lesser extent, triclopyr. Procymidone was much less mobile but was very persistent. Atrazine, terbuthylazine, and 2,4-D showed low mobility and persistence at the study site. Bromide and hexazinone were observed in the underlying groundwater following winter recharge and an intensive irrigation period at the end of the study. There was slight adsorption of bromide by the allophane in the soil. The combination of both soil sampling and suction cups gave a more complete description of the leaching process and the distribution of the compound through the profile than either technique by itself. This combination is worthwhile and gives complementary information providing the strengths and limitations of both techniques are appreciated.

Seven pesticides were applied to an allophanic silt loam along with a bromide tracer and their concentrations in soil and water monitored over a 2-year period. Inverse modelling was carried out using GLEAMS, LEACHM, and HYDRUS-2D to derive field-based mobility and degradation parameters. Hexazinone and procymidone were more mobile and more persistent than most literature values would suggest, whereas picloram and triclopyr were much less mobile but more persistent. The greater mobility for hexazinone, a weak base, and the reduced mobility of picloram and triclopyr, weak acids, are consistent with the effects of allophane. Mobility values for 2,4-D, atrazine, and terbuthylazine could not be determined with confidence from experimental results, but both atrazine and terbuthylazine appeared less persistent, and 2,4-D more persistent than literature values. A fourth model, SPASMO, currently under development, was used as an independent test of the optimised parameters. It performed well for the soil water concentrations but tended to overestimate the observed soil concentrations using the derived parameters. HYDRUS-2D simulations of bromide and hexazinone concentrations in the groundwater gave a good fit to observed data from 3 monitoring wells following a large recharge pulse.