Ranchers are increasingly expected to manage grasslands for forage production and native biodiversity enhancement goals. However, longstanding relationships between grazing and plant species are often understudied because elucidating effects of grazing absence and presence often requires experimental opportunities that are difficult to establish, such as the introduction of grazing to long-term ungrazed pastures. Addressing this knowledge gap is critical for heterogeneous landscapes where site-specific properties might interact with grazing effects to ultimately structure plant communities. We conducted vegetation surveys for 3 years after grazing was reintroduced to an annual California grassland that was not grazed for more than 60 years. We investigated how grazing affected plant communities in terms of cover and richness of native and invasive species and how topographic sites of summit, backslope and toeslope altered these relationships. The plant communities were affected by the independent effects of grazing, site and year. Across years, native cover was 39% greater in grazed plots compared with ungrazed plots. Native species richness was slightly lower in ungrazed compared with grazed plots for toeslope sites relative to the other topographic positions. Invasive species cover was 17% lower in grazed plots compared with ungrazed plots and no predictors were found to contribute to significant differences across plots. Although we generally did not find expected relationships between site and plant response to grazing, this work demonstrates how managers can use livestock to quickly modify plant communities in areas with a long history of grazing absence.