Variation in characteristics and conservation values of plant communities on abandoned agricultural lands with and without fires

QUESTION: Which plant community characteristics most consistently vary with prior land use? LOCATION: Oak Openings Preserve, Ohio, USA. METHODS: In 2020, we measured plant communities on 22 sites encompassing three prior land uses: (a) contemporary Quercus forests continuously forested since before the 1930s; (b) former agricultural lands, under cultivation in the 1930s and abandoned 70–80 years before our study and that were unburned; and (c) former agricultural lands also abandoned 70–80 years earlier and that were burned in 2013 prescribed fires. Among the three land uses, we compared 12 plant community variables, categorized into four groups characterizing different sets of community properties (structure, diversity, composition, and conservation values). RESULTS: A multivariate combination of the 12 community variables differed among all three land uses, each of the four categories of variables differed between at least two land uses, and eight of 12 individual variables differed between at least two land uses. Community structure displayed a gradient of decreasing tree canopy and understorey plant cover and increasing lichen–moss cover from continuously forested to unburned and burned formerly cultivated sites. Understorey plant species richness did not vary with land use, but species diversity was higher on cultivated sites, irrespective of fire history. While all three land uses now contain Quercus tree overstories, each land use exhibited unique understorey species composition. Only formerly cultivated sites contained state‐listed endangered species. Cultivated sites had a lower proportion of native species cover than did continuously forested sites, but native species still comprised over 90% of the plant cover on cultivated sites. Compared with continuous forests, the soil organic layer (O horizon) was five times thinner and loss on ignition in the 0–15 cm mineral soil 49% (unburned) and 66% (burned) lower on previously cultivated sites. CONCLUSIONS: Community characteristics varied in their sensitivity to land use history, with species composition among the most sensitive to prior land use. A dense sapling layer of Acer rubrum forming in long‐unburned Quercus forests across much of eastern North America was absent in our study on formerly cultivated sites, possibly due to dry, infertile soil. From a restoration and conservation standpoint of the study region's pre‐settlement, frequently burned and open savannas and woodlands, previously cultivated sites may be easier to keep open via prescribed fire. Moreover, perhaps precisely because much of the flora historically developed in open savannas and woodlands which previously cultivated sites on the contemporary landscape most closely mimicked, previously cultivated sites presently contain higher species diversity and more conservation priority species than do continuously forested sites.