Exotic weed species can affect native populations, communities, and ecosystem processes.  Land managers must often choose early control and monitoring programs without knowing the immediate threat that a potential invader poses.  I conducted a two and a half year study of herbicide application on a patchily distributed population of Lespedeza cuneata, a perennial legume of Asian origin.  This experiment was designed to address the basic questions: 1) To what degree will herbicide affect spread and how variable are the outcomes? 2) Does collecting more detailed data provide additional biological insight?  During each annual transition L. cuneata occupancy increased in ~63% of the untreated plots.   While occupancy decreased in 73% of the spot sprayed plots between 2002 and 2003, only 27% of the sprayed plots experienced a decrease between 2003 and 2004.  Binary and density-based models made qualitatively similar predictions of occupancy at the one meter scale.  Nearly all models indicated that the odds of colonization are low, and the odds of persistence are high.  However, the density-based models indicated that local and/or neighborhood abundance must be included in estimates of spread.  These results suggest that though herbicidal control could be variable, failure to continually treat the species may lead to its steady spread.  They also indicate that including abundance in predictions of spread may prevent managers from underestimating future occupancy and the resources required to effectively manage the population.