Efficacy and non-target effects of herbicides in foothills grassland restoration are short-lived

Questions: Selective herbicides are frequently used in ecological restoration to control invasive non-native forbs and recover plant communities. However, the long-term efficacy of this practice, its non-target effects on native plants, and its role in facilitating secondary invasions are not well understood. Similarly, little is known about the extent to which herbicide drift may affect native plant communities. Location: Foothills grasslands of Montana, USA. Methods: We conducted a 6-year experiment to investigate changes in the abundance of a target invasive plant, knapweed (Centaurea stoebe subsp. micranthos) and plant community structure in response to the herbicides Tordon® (picloram) and Milestone® (aminopyralid), applied at a recommended rate and a diluted rate that simulated drift. Results: Knapweed cover and the richness of native and non-native forb species declined in the first 3 years in response to treatment at recommended rates, but not drift rates. Secondary invasion by non-native monocots was significant but weak. The cover of native forbs and the cover and richness of native monocots did not differ among treatments but changed significantly with the year. Surprisingly, 6 years after treatments, there were no differences among treatments in the cover of the target invasive plant or community structure. Conclusions: Our results demonstrate that the efficacy and non-target effects of herbicides in grassland restoration can be short-lived and idiosyncratic because of year effects. Restoration of knapweed invasions might require other active interventions, such as seeding or repeated spraying. Our study supports previous calls for long-term monitoring of herbicides application in ecological restoration.