Spring precipitation amount and timing predict restoration success in a semi‐arid ecosystem

Abstract Successful restoration of semi‐arid ecosystems remains a challenging problem. Many studies have tried to elucidate the factors that limit seedling establishment so that mitigating actions can be taken during restoration. However, most studies have been plot‐based trials on individual sites that encompass only 1 or 2 years of meteorological variation. This limits the inferences that can be drawn. To move beyond short‐term research, we examined seedling establishment after the first growing season on 48 public lands sites undergoing restoration in the sagebrush steppe of the northern Great Basin over the course of 5 years (2016–2020), while simultaneously accounting for a broad suite of biotic (five), abiotic site (six), management (five) and weather variables (29). All these variables have been previously identified as influencing successful restoration. Using least absolute shrinkage and selection operator regression and a random forest algorithm, we identified nine variables associated with surviving seedlings at the end of the first growing season. Seedling establishment averaged 9.6 (±10.2 SD) individuals m −2 . Precipitation variables accounted for 75% of the explained variation in seedling survival. April precipitation was the largest contributor and accounted for 27% of explained variation. Average length of dry period in February and number of rainfall events >5 mm March–May were also important predictors of survival, accounting for 13% and 10% of the explained variation, respectively. Only one biotic variable (annual grass density) and one management variable (drill seeding) accounted for more than 5% of the explained variance by the model. While temperature may be important in many models of seedling establishment, water availability during the emergence period (February–April) appeared to play the most important role determining the number of seedlings surviving at the end of the first growing season in this study. Synthesis and applications . The importance of April precipitation in predicting seedling establishment densities makes it a variable that should be included in restoration success forecasting models. Similarly, since dry periods and precipitation frequency combined explain nearly as much variation in establishment, efforts should be made to improve seasonal forecasting to include predictions of frequency and event size.