Every restoration is unique: testing year effects and site effects as drivers of initial restoration trajectories

Summary The outcomes of restoration efforts are contingent on the specifics of the restoration practices utilized, but also on uncontrolled contingencies such as site effects and year effects. Although restoration practitioners have long been aware that the successes of their projects vary from site to site and from year to year, there have been few direct experimental tests of these contingencies. We established grassland restoration plots identically across three sites in northern California, in each of four establishment years (for 12 site‐year combinations). The resulting plant communities differed significantly across sites and across establishment years. As a consequence of these community differences, there were ‘forb years’ and ‘grass years’, although these sometimes differed among sites. Multivariate analysis identified mean annual temperature and total precipitation as likely drivers of some of these differences. Synthesis and applications . Our results not only confirm the idiosyncratic nature of the results of restoration efforts (and ecological experiments in general) but also demonstrate that some of this variation can potentially be related to measurable environmental conditions. Understanding the drivers of this variability can ultimately aid restoration practitioners by allowing them to focus restoration efforts on years and sites most likely to yield desired outcomes.