Introduced exotic plants can drive ecosystem change. We studied invasion and establishment of Brassica tournefortii (African mustard), a noxious weed, in the Chemehuevi Valley, western Sonoran Desert, California. We used long-term data sets of photographs, transects for biomass of annual plants, and densities of African mustard collected at irregular intervals between 1979 and 2009. We suggest that African mustard may have been present in low numbers along the main route of travel, a highway, in the late 1970s; invaded the valley along a major axial valley ephemeral stream channel and the highway; and by 2009, colonized 22 km into the eastern part of the valley. We developed predictive models for invasibility and establishment of African mustard. Both during the initial invasion and after establishment, significant predictor variables of African mustard densities were surficial geology, proximity to the highway and axial valley ephemeral stream channel, and number of small ephemeral stream channels. The axial valley ephemeral stream channel was the most vulnerable of the variables to invasions. Overall, African mustard rapidly colonized and quickly became established in naturally disturbed areas, such as stream channels, where geological surfaces were young and soils were weakly developed. Older geological surfaces (e.g., desert pavements with soils 140,000 to 300,000 years old) were less vulnerable. Microhabitats also influenced densities of African mustard, with densities higher under shrubs than in the interspaces. As African mustard became established, the proportional biomass of native winter annual plants declined. Early control is important because African mustard can colonize and become well established across a valley in 20 yr.
Nomenclature: Brassica tournefortii Gouan BRSTO.
Management Implications: Our models of African mustard invasion and establishment in the western Sonoran Desert have broad application for designing surveillance and monitoring systems in other arid and semiarid lands in the American West. With this study we now know that topography (major and small ephemeral streams or “washes” in valleys) and ages of geological deposits and soils are important predictors of sites for African mustard invasion. In addition to roads, African mustard is most likely to invade and colonize areas through active major and minor ephemeral stream channels, especially where these features are numerous on the landscape or intersect highways and roads. These features are highly vulnerable because they are naturally disturbed and composed of young geological deposits (1 to 7 yr old). Less vulnerable are older deposits associated with desert pavements (20,000 to 300,000 yr old), and least vulnerable are the oldest geological deposits (140,000 to 300,000 yr) with well-developed desert pavements. Weed control specialists and land managers can use existing regional maps of topography, surficial geology, and landforms, coupled with aerial imagery, to design surveillance systems that emphasize ephemeral stream channels and young geological surfaces. Aerial views of desert ephemeral stream channels suitable for mapping can be obtained from sources such as the National Agricultural Imagery Program. Monitoring for evidence of arrival and early invasion can be achieved by walking transects several kilometers in length, both up and down stream channels, and where such channels cross or parallel roads or other disturbed areas. When the mustard is located, weed control specialists can take action to eradicate or limit spread with herbicides, which can be more effective and less costly than physical methods. Timing of herbicide application, i.e., at an early phenological stage, may determine effectiveness in controlling not only the mustard but also other exotic plant species. Herbicides that do not negatively affect native annuals (based on limi