Rates, patterns, and drivers of tree reinvasion 15 years after large‐scale meadow‐restoration treatments

Tree encroachment threatens the biodiversity and ecological functioning of grasslands worldwide. Reversing effects of encroachment requires not only removing trees, but limiting subsequent invasions, which can stall grassland reassembly. We quantified rates and patterns of conifer reinvasion of mountain grasslands (meadows) in western Oregon, 7 and 15 years after experimental tree removal with or without burning. We assessed frequency (percentage of 100 m(2) subplots) and density of reinvading trees in six 1-ha plots as a function of burn treatment, abundance in the adjacent forest (a proxy for seed availability), and distance to or orientation of forest edges. After 15 years, frequency and density of colonists varied widely among plots (24-57% of subplots and 60-250 trees/ha, respectively), but burning had no effect on either measure. Abies grandis, which dominated adjacent forests, was the principal invader. Colonization by less-common Pinus contorta and A. procera was correlated with species' density and basal area in adjacent forests, suggesting that establishment was seed limited. Annualized rates of invasion were similar between early and late sampling intervals (9-10/trees/ha). Establishment was generally greater along the forested edges of plots (0-10 m) than in the core (10-50 m). Prominence of shade-tolerant A. grandis along northerly-facing edges suggests strong microclimatic controls on establishment. In sum, tree removal, with or without prescribed fire, is successful in reversing effects of conifer encroachment. Despite abundant seed rain, reinvasion is slow and spatially restricted. Periodic hand removal of colonizing trees, while they are small, can be done efficiently and economically.