Effects of herbicide spray drift and fertilizer overspread on selected species of woodland ground flora: comparison between short-term and long-term impact assessments and field surveys.
Although application of fertilizers and pesticides to farmland is known to impact non-target organisms in adjacent semi-natural habitats, little is known of the impact on woodland understorey communities, which have considerable conservation significance in the arable landscapes of the UK and elsewhere in Europe. This study investigated the impacts of agriohemicals on several plant species typically found in ancient semi-natural communities. Six species of woodland plants were exposed to the herbicide glyphosate at concentrations equivalent to those measured in spray drift trials (1-25% of the full application rate) in short-term greenhouse and long-term field experiments. Fertilizer was applied to half of the plants at a concentration likely to be seen in overspread situations (70 kg ha-1 year-1). The distribution of the same species was surveyed in woodland margins adjacent to fields with high, medium and low agrochemical inputs. In both greenhouse and field experiments, herbicide treatments at drift concentrations caused increased mortality, reduced biomass and reduced fecundity in all species, although to differing degrees. The threshold of sensitivity to glyphosate was as low as 1% of the median field application rate for the most sensitive species. Fertilizer treatment affected resource partitioning in Carex remota and Galium odoratum and reduced the fecundity of G. odoratum. In field surveys, the abundance of the most sensitive species was highest in woodland margins adjacent to fields with low agrichemical inputs and lowest beside high-input farmland. These differences were seen at least 4 m into woodland margins but were not found at 10 m. There was considerable agreement between the short-term greenhouse and long-term field experiments, although long-term impacts may have been underestimated in the former. Adverse impacts of pesticide drift have been demonstrated to affect a variety of plants of woodland margins bordering farmland. Differences in the distribution of species most sensitive to herbicide extend to approximately 4 m into woodland margins. We recommend the adoption of no-spray buffer zones of approximately 5 m to protect the majority of woodland species from the impacts of agrichemicals applied to adjacent land.