Non-indigenous invasive species are a major threat to native species diversity and ecosystem function and have been called the single worst threat of natural disaster of this century. Eragrostis lehmanniana Nees (Lehmann lovegrass), a tufted perennial bunchgrass native to southern Africa, is one such problematic species in Arizona, USA. This dissertation research is a mix of predictive modeling and field experiments designed to inform management decisions based on greater understanding of this nonnative species, with emphasis on the potential for spread and the impacts of removal.The modeling studies in this dissertation aimed to predict the potential distribution of E. lehmanniana in the southwestern United States under current and potential future climate conditions. The first portion of study addressed a common assumption in predictive modeling of nonnative species: data from the species' native range are necessary to accurately predict the potential distribution in the invaded range. The second portion of this study predicted the distribution of E. lehmanniana under 28 different climate change scenarios. Results showed the distribution of E. lehmanniana progressively shrinking in the southeastern and northwestern portions of the state and increasing in the northeastern portion of the state with increasing temperatures and precipitation. Key shifts occurred under scenarios with increases in summer and winter precipitation of 30% or more, and increases in summer maximum and winter minimum temperatures of at least 2oC.The field experiment served as a pre-eradication assessment for E. lehmanniana and indicates how semi-desert grassland communities in southeastern Arizona may respond to the removal of this species. This study suggested that plant community response to removal of an introduced species is mediated by precipitation variability (timing and amount), local site history, and edaphic conditions. The response observed on a site previously farmed for decades was to subsequently become dominated by other nonnative annual species. However, the two other sites with histories of livestock grazing responded more predictably to the removal, with an increase in annual ruderal species (2 to 10 times the amount of annual cover recorded on control plots).