Active restoration after three decades: Seed addition increases native dominance compared to landscape‐scale secondary succession

Abstract Active restoration often aims to accelerate ecosystem recovery. However, active restoration may not be worthwhile if its effects are overwhelmed by changes that occur passively. Moreover, it can be challenging to separate the effects of passive processes, such as dispersal and natural succession, from active restoration efforts. We assess the 24‐year impact of actively restoring a Minnesota old‐field grassland via seed addition of native tallgrass prairie species. We compared the abundance of four functional plant groups in actively restored plots against abundances in three reference classes: (1) unrestored plots undergoing passive recovery within the same old field, (2) passively recovering plots in two nearby old fields of similar age and (3) a chronosequence of 21 old fields within the same landscape. Active restoration led to a higher abundance of native grasses and forbs in the 36 m 2 treatment plots. Seed addition was more effective if the original vegetation was first removed using herbicide, burning and tilling. However, long‐term conclusions about the efficacy of active restoration varied widely depending on the choice of reference class. In our small‐scale restoration experiment, native abundance was similarly high in both the actively restored and reference plots after 24 years, suggesting either (1) passive recovery or (2) local dispersal of native species from nearby treatment plots (i.e. cross‐contamination). In contrast, a comparison with two nearby reference fields suggested active restoration resulted in much higher native abundance relative to passive recovery. A smaller, positive effect was detected when we compared actively restored plots to the chronosequence of old fields. In the chronosequence, many passively recovering old fields had transitioned to native grass dominance naturally, although active restoration appeared to increase native forb abundance. Synthesis and applications : Our findings highlight the importance of using scale‐appropriate references for assessing the efficacy and need for active restoration. Comparing actively restored plots with the surrounding landscape, we found that active restoration and passive recovery led to similar plant communities after 24 years. Because local dispersal from actively restored sites can nearby references, caution should be exercised when evaluating long‐term restoration projects using only small‐scale experiments.