The impact of herbicide tolerant oilseed rape in selected agro-ecosystems

Simpson, Euan C (2002) The impact of herbicide tolerant oilseed rape in selected agro-ecosystems. PhD thesis, University of Glasgow.

Abstract

A range of field studies of cross pollination between herbicide tolerant and conventional oilseed rape crops and plots were conducted to demonstrate the effects of variety, distance, pollen source and sink size and intervening crop on levels of outcrossing. Experiments investigating the influence of variety on outcrossing showed that hybrid oilseed rape varieties containing high proportions of male sterile plants (varietal associations) were pollinated at higher frequencies than standard fully fertile varieties. Studies conducted using various sizes of genetically modified herbicide tolerant (GMHT) pollen sources showed that small GMHT feral populations cross pollinated with crops in close proximity and the levels of contamination obtained depended on the genotype of the conventional crop. Long range cross pollination of male sterile oilseed rape receptor plots showed that pollination events were measurable at up to 600m from the large GMHT pollen source. Results from a study of cross pollination between mixed populations of GMHT oilseed rape plants and conventional varieties also demonstrated that a hybrid rape type (varietal association) was cross pollinated at considerably higher levels than an open pollinated and a fully restored hybrid variety. Evidence gathered in the experiment suggested that, over a wide range of initial GMHT contamination rates, the final proportion of GMHT seed in the total population was a constant fraction of the initial contamination rate. Outcrossing data was used to compare negative exponential and inverse power law models for their fit to describe the observed relationship between cross pollination and distance from source. Results showed that the inverse power law provided a better fit of the data. This demonstrated that dispersal described by the inverse power law was more likely to lead to cross pollination at both near and large distances from the pollen source compared to the negative exponential model. The consequences of the likely ecological behaviour of GMHT traits resulting from the dispersal curves for regulation and risk assessment are discussed. The effect of the herbicides used in herbicide tolerant and conventional oilseed rape on weed populations were compared in a single season. Results suggested that the herbicides have different activity spectra thus resulting in a variety of surviving weed species in HT treatments. The change in active ingredient and the timing of herbicide application in HT winter oilseed rape crops will likely cause a change in the weed species that are being controlled or those that escape treatment. Levels of weed biomass recorded prior to harvest of the oilseed rape crop showed that there may be differences between treatments in terms of the quantity of seed returned to the seedbank from the range of weed species present. Limited data on the behaviour of herbicide tolerant volunteers showed that single and putative double-tolerant plants were as susceptible as conventional oilseed rape volunteers to normal selective herbicides used in cereal crops. Data from a number of elements of the studies on GM contamination rates, weed control, and seed bank estimates were used to develop a simple population projection model. The model used a Markov process to examine the fate of volunteer and feral populations of oilseed rape comprising a mixture of conventional and herbicide tolerant types. Results from the model indicated that the prevalence of the GMHT trait in the weed or feral population was more sensitive to the efficacy of control practices used in the rotation than the levels of cross-pollination and competition between the herbicide tolerant and conventional varieties in the mixed population. Thus, although the cross-pollination studies suggested that varietal associations are more likely to be cross-pollinated by GM pollen than fully fertile varieties, the projection model suggested that resulting differences in the prevalence of the GM trait in volunteer and feral populations may not be very large. Population projections from the model are compared with results from other modelling studies which have used more complex simulation approaches.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Additional Information: Adviser: George Marshall
• Keywords: Agronomy
• Date of Award: 2002
• Depositing User: Enlighten Team
• Unique ID: glathesis:2002-72122
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 24 May 2019 15:11
• Last Modified: 24 May 2019 15:11
• URI: https://theses.gla.ac.uk/id/eprint/72122