A Low Cost Method for Glyphosate Analysis, and Site Investigation and Modelling of Glyphosate Fate and Transport from Genetically Modified Canola Farmland in Parkes, NSW, Australia

Since 2000, the use of a glyphosate, a broad-spectrum herbicide, has increased dramatically due to the cultivation of genetically modified crops in Australia. The research conducted for this thesis analysed glyphosate in environmental samples obtained from an agricultural area in Parkes, NSW, and its fate and transport in an agricultural system. Firstly, a low-cost analytical method for glyphosate analysis was established in soils and waters, to avoid the need for much more expensive analytical methods such as High Performance Liquid Chromatography. The method was calibrated against a commercially available enzyme-linked immunosorbent assay (ELISA) method, for the analysis of glyphosate which involves derivatisation with 9-Fluorenylmethyl chloroformate (FMOC-chloride) and measurement of the emissions acquisition at 310 nm with excitation wavelengths of 268 nm. Using linear regression techniques, the calibration curves showed linearity for ranges from 10 µg/L to 25000 µg/L and 10 µg/L to 1,000 µg/L (Models I and III respectively) with good reproducibility (recovery 108.3 µ20.8; R2=0.997 and 115.230 µ25.852; R2=0.998 respectively). Secondly, a site investigation was conducted of farmland cultivated with GM canola in the Parkes region, New South Wales, Australia, which involved the sampling of surface waters and soils and glyphosate analysis. The sampled soils are loams under the USDA system (clay 13.8-15.8%, silt 39-43%, sand 41.2-47.2%) and were collected according to the Australian National Environmental Protection Measure sampling protocol. The found glyphosate levels ranged between 10 and 67 µg/L in waters and 0.10 and 0.575 mg/kg in soils, with aqueous levels lying below Australian and international drinking water guidelines. However, some exceeded the Canadian Freshwater Guideline for the Protection of Aquatic Life of 65 µg/L and most exceeded the South African chronic guideline of 0.3 µg/L. Thirdly, glyphosate sorption isotherms were constructed by batch tests on several soils from the study site and fitted with linear and non-linear Langmuir, Freundlich and Redlich-Peterson isotherms and desorption tests performed using a 0.1M KH2PO4 solution. Column-leaching experiments were also conducted on glyphosate-dosed and were modelled using the one-dimensional transport model HYDRUS-1D with sorption and degradation. Finally, the findings are incorporated into a conceptual model of glyphosate transport pathways and environmental receptors