A central goal of ecosystem restoration is to promote diverse, native-dominated plant communities. However, restoration outcomes can be highly variable. One cause of this variation may be the decisions made during the seed mix design process, such as choosing the number of species to include (sown diversity), or the number of locations each species should be sourced from (source diversity, manipulated to affect genetic diversity). The effects that seed mixes have on plant communities may be further modified by other factors at the restoration site, including edge proximity and consumer pressure, which are also likely to vary over time. However, few studies have evaluated these seed mix attributes together, and none have done so under realistic field conditions. To address this research need, we conducted a prairie restoration experiment where two aspects of a seed mix (sown diversity and source diversity), edge proximity, and vertebrate consumer access were manipulated across 12 replicate fields. We found that when seed mix design impacted plant community structure, these effects were dependent on consumer access or edge proximity and were more prominent after one vs. five growing seasons. Low seed source diversity plots had more sown species than high source diversity ones, but only when consumers had access. Similarly, low species diversity plots had higher richness and cover of species included in both the low and high species diversity mixes, but this effect weakened over time. Additionally, plots with high species-diversity were buffered from the typically detrimental effects of edges and consumers, although this did not always result in greater sown species abundance. Unexpectedly, plots with the most sown species were those sown with either low source diversity or low species diversity seed mixes, perhaps due to lower seeding rates of reliably establishing species. Our results illustrate how the influences of seed mix design on restored plant communities can be highly contingent on factors like edges, consumers, and time.

Data for this manuscript was collected in 2015 and 2021 from a long-term prairie experiment at Kellogg Biological Station where two aspects of a seed mix (sown diversity and source diversity), edge proximity, and vertebrate consumer access were manipulated across 12 replicate fields. Data was analyzed with R studio using provided code.