Frequency‐dependent herbicide effects on rangeland plants and soils exceed those of the invader

Invasive species management is a significant challenge and often involves herbicides. Herbicides, however, can have significant nontarget effects and often need to be reapplied to maintain control, which may amplify negative effects. Effects on plant communities are well documented, but less is known about herbicide effects on rangeland soils with respect to application frequency. Given strengthening plant community effects with repeated application, we should see concurrent changes in soil microbiomes and soil functioning. To explore how application frequency affects rangeland plants and soils, we focused on leafy spurge ( Euphorbia esula L.) invasion. Leafy spurge is invasive in North America and causes widespread changes in rangeland ecosystems, with herbicides often the control method of choice. We applied a common broadleaf‐specific herbicide containing aminocyclopyrachlor and metsulfuron‐methyl (Navius FLEX; Envu) either never, once, or in two consecutive years in both invaded and uninvaded areas. We then measured plant biomass and composition, and soil microbial composition, carbon, and nitrogen for 2 or 3 years. One application reduced leafy spurge nearly 90% for two growing seasons, declining to 60% by the third. One application also caused approximately 50% declines in forb biomass and broadleaf species richness, and a 300% increase in nitrate, with limited other soil effects. By year 3, the plots recovered; however, there was also a 45% increase in secondary invasion by exotic grasses. A second application worsened most plant community effects and caused an 85% loss of shrubs. Further, the second application caused changes in the soil microbiome, with a 22% decline in fungal abundances and increased stress in bacteria. The changes in the microbiome are likely due to the loss of good mycorrhizal hosts and belowground carbon allocation, which is supported by a concurrent 25% decline in labile carbon. Conversely, we found no significant differences in plant and soil properties between invaded and uninvaded areas. Herbicide can thus have frequency‐dependent detrimental effects on rangeland ecosystems that become more pronounced with a second application. Overuse of herbicides for invasive control thus has the potential to significantly alter biodiversity and ecosystem structure and may have worse effects than the invasive species themselves.