Transformations and mobility of triazine metabolites in two New Zealand soils

Degradation and sorption parameters were determined for two triazine herbicides, atrazine, and simazine, and their two metabolites, desethylatrazine (DEA), and desisopropylatrazine (DIA) in a Horotiu silt loam (7.3% organic carbon, 12% allophane) and a Te Awa silt loam (2.17% organic carbon). Inverse modelling was carried out using PEST (a parameter optimisation package) and LEACHM to derive field-based sorption and degradation parameters. Simulations using LEACHM were used to estimate pesticide and metabolite distribution through the Horotiu and Te Awa soils. Laboratory-determined half-lives of the metabolites ranged between 31-32 days for DEA, and 24-34 days for DIA in both soils, similar to the half-lives of atrazine and simazine (20 days and 32-37 days, respectively). There was good agreement between the Freundlich sorption constants (Kf = 3.38, 3.59, 1.70, 2.13 for Horotiu, and Kf = 1.17, 1.43, 0.69, 1.22 for Te Awa, for atrazine, simazine, DEA, and DIA respectively) and the linear pesticide distribution constants (Kd = 3.96, 4.90, 1.97, 3.03 for Horotiu, and Kd = 1.13, 1.46, 0.70, 1.31 for Te Awa, for atrazine, simazine, DEA, and DIA respectively). The similarities between metabolites and their parent compounds in sorption coefficients and degradation rates indicate that DEA and DIA need to be considered with atrazine and simazine when the groundwater contamination potential of triazine pesticides are examined. The inverse modelling of the field experiments resulted in only a few parameters which were estimated with confidence. Sorption and degradation values for DEA could not be determined with confidence from the Horotiu field results. The mobility and sorption of DEA and DIA could not be determined with any confidence in the Te Awa soil. However, atrazine and simazine were estimated to be less mobile and less persistent than literature values. The LEACHM simulations using the laboratory-derived degradation and sorption parameters were compared with the measured field observations from both soils. The pesticide and metabolite distribution was overestimated in the topsoil (