A regional land transformation is underway in the Sonoran Desert of southwestern North America as a result of the conversion of native rangeland to exotic pasture. In northwestern Sonora, Mexico the process involves clearing native vegetation for cultivation of buffelgrass (Pennisetum ciliare). Southern African buffelgrass was introduced to Sonora through the United States in the 1950s with generous support from the Mexican federal and Sonoran state governments. The ascendance of buffelgrass as a range management tool in Sonora has been conditioned by an international political economy of beef production. However, land use decisions regarding buffelgrass are also conditioned by factors internal to the ranch household. This research examines the expansion of buffelgrass in the Sonoran Desert, addressing its extent and drivers. Through the use of systematic interviews with ranchers, key informant interviews with government officials, and an examination of northern Mexico's cattle ranching history and policy, the dissertation documents why buffelgrass has spread as a policy program and management choice. This part of the work addresses a "structure-agency debate" in human-environment geography. Next the research turns to landscape impacts of buffelgrass cultivation, through vegetation plot and transect sampling. The extent, cover, and density of buffelgrass inside and outside fenced pastures are examined, confirming the hypothesis that disturbance facilitates invasion from pastures onto surrounding lands. Finally, the research employs novel methods of remote sensing and geographic information science using a 1973-2006 time series of Landsat imagery to characterize the patterns and temporal trajectories of land change by buffelgrass across the site. Object-based image processing techniques are combined with traditional maximum likelihood techniques and classification tree analysis to address the difficult task of distinguishing buffelgrass from other prevalent land covers. Together, these topics and methods constitute an application of "land change science," which merges traditional methods from the social, biophysical, and geospatial sciences to arrive at an integrated characterization and explanation of landscape change in coupled human-environment systems.