1. Seed-based restoration is one of the most difficult challenges for dryland restoration. Identifying environmental conditions that drive variation in seed and seedling mortality across similar restoration efforts could increase understanding of when and where restoration outcomes are likely to be favorable and identify new tools and strategies to improve outcomes. 2. We asked how variation in a suite of environmental predictors influenced germination, emergence, seedling establishment, and juvenile survival of four commonly sown perennial grass species across 33 seeding experiments distributed over a ~160,000-km2 area of the Great Basin, a cold desert system in the western United States. 3. Across experiments, we observed wide variation in the rates of four demographic transitions and wide variation in environmental conditions experienced by plants at each stage. For all species, higher precipitation during the first 30 days following seeding was associated with an increase in germination. Conversely, higher soil temperature over this same time period was associated with a significant decrease in germination and emergence and soil temperature was associated with a substantial portion of the variation in germination and emergence probabilities observed across our seeding experiments. 4. Within the range of precipitation variation observed, we were unable to detect a significant relationship between seedling establishment the first growing season and cumulative precipitation the first year, precipitation during the first spring growing season, or annual climatic water deficit (CWD) the first year. Higher CWD the second growing season reduced seedling survival over that time period. 5. Synthesis and application. Our results show higher soil temperature negatively impacts grass seedling recruitment. Our results can be combined with seasonal and subseasonal temperature forecasts to improve restoration decision-making. These results also suggest climate warming will make restoration even more difficult, with our model estimates suggesting the 2°C increase in temperature expected in the Great Basin over the coming decades will decease germination and emergence by about 30%. Lastly, while our field-based approach provided insight into short-term drivers of mortality, it did not provide insight into drivers of longer-term survival, suggesting a need to develop process-based approaches for predicting long-term restoration outcomes.