1. Worldwide, grasslands are becoming shrublands/forests. In North America, eastern red cedar (Juniperus virginiana) often colonizes prairies. Habitat management can focus on woody removal, but we often lack long-term data on whether removal leads to population recovery of herbaceous plants without seeding. 2. We undertook a long-term study (17 years) of numbers of the rare annual plant Agalinis auriculata in a gridwork of 100 m2 plots in adjacent prairie and oldfield sites in Kansas, USA. We collected data before and after removal of Juniperus virginiana at the prairie. 3. Plant population sizes were highly variable at both sites and over time. High numbers of plants in a plot one year were often followed by low numbers the following year, suggesting negative density-dependence. Plant numbers were lowest with extensive woody cover and with low precipitation. After woody plant removal, A. auriculata increased dramatically in abundance and occupancy in most years; increases were also seen at the oldfield, suggesting later survey years were overall more favorable. 4. Synthesis and applications: Removal of woody plants led to increased numbers of a rare annual prairie plant, without seeding. Multiple years of data were essential for interpretation given extreme temporal variability in numbers. The largest prairie population was seven years following tree removal, showing that positive effects of management can last this long. This species also fared well in oldfield habitat, suggesting restoration opportunities. Given that land managers are busy, time-efficient field methods and data analysis approaches such as ours offer advantages. In addition to general linear models, we suggest Rank Occupancy Abundance Profiles (ROAPs), a simple-to-use data visualization and analysis method. Creation of ROAPs for sites before and after habitat management helps reveal the degree to which plant populations are responding to management with changes in local density, changes in occupancy, or both.