Evaluating the Role of Habitat Quality on Establishment of GM Agrostis stolonifera Plants in Non-agronomic Settings

Abstract

We compared soil chemistry and plant community data at non agronomic mesic locations that either did or did not contain genetically modified (GM) Agrostis stolonifera. The best two-variable logistic regression model included soil Mn content and A. stolonifera cover and explained 90% of the variance in the probability of a site having GM A. stolonifera. Inclusion of NH₄ as a third predictor variable increased the variance explained by the logistic model to 100%. Soils at GM locations were characterized by significantly lower (P < 0.05) Mn, A. stolonifera cover, and NH₄. Pairwise comparisons indicated that sites in which the GM plants became established had a significantly higher % of bare ground and significantly lower A. stolonifera cover, Mn, organic matter, and carbon (P < 0.05). The pH of soil at GM plant locations varied from 5.9 to 9.5. Our results suggest potential roles of soil disturbance and nutrient status in the establishment of Agrostis in mesic habitats. Additional research is needed to evaluate the ecological consequences of gene flow of GM Agrostis to non-agronomic plant communities.