Unprecedented modern rates of warming are expected to advance boreal forest into Arctic tundra1, thereby reducing albedo2–4, altering carbon cycling4 and further changing climate1–4, yet the patterns and processes of this biome shift remain unclear5. Climate warming, required for previous boreal advances6–17, is not sufficient by itself for modern range expansion of conifers forming forest–tundra ecotones5,12–15,17–20. No high-latitude population of conifers, the dominant North American Arctic treeline taxon, has previously been documented5 advancing at rates following the last glacial maximum (LGM)6–8. Here we describe a population of white spruce (Picea glauca) advancing at post-LGM rates7 across an Arctic basin distant from established treelines and provide evidence of mechanisms sustaining the advance. The population doubles each decade, with exponential radial growth in the main stems of individual trees correlating positively with July air temperature. Lateral branches in adults and terminal leaders in large juveniles grow almost twice as fast as those at established treelines. We conclude that surpassing temperature thresholds1,6–17, together with winter winds facilitating long-distance dispersal, deeper snowpack and increased soil nutrient availability promoting recruitment and growth, provides sufficient conditions for boreal forest advance. These observations enable forecast modelling with important insights into the environmental conditions converting tundra into forest.

Understanding the key mechanisms that control northern treelines is important to accurately predict biome shifts and terrestrial feedbacks to climate. At a global scale, it has long been observed that elevational and latitudinal treelines occur at similar mean growing season air temperature (GSAT) isotherms, inspiring the growth limitation hypothesis (GLH) that cold GSAT limits aboveground growth of treeline trees, with mean treeline GSAT ~6–7°C. Treelines with mean GSAT warmer than 6–7°C may indicate other limiting factors. Many treelines globally are not advancing despite warming, and other climate variables are rarely considered at broad scales. Our goals were to test whether current boreal treelines in northern Alaska correspond with the GLH isotherm, determine which environmental factors are most predictive of treeline presence, and identify areas beyond the current treeline where advance is most likely. We digitized ~12 400 km of treelines (>26 K points) and computed seasonal climate variables across northern Alaska. We then built a generalized additive model predicting treeline presence to identify key factors determining treeline. Two metrics of mean GSAT at Alaska's northern treelines were consistently warmer than the 6–7°C isotherm (means of 8.5°C and 9.3°C), indicating that direct physiological limitation from low GSAT is unlikely to explain the position of treelines in northern Alaska. Our final model included cumulative growing degree‐days and near‐surface (≤1 m) permafrost probability, which together may represent the importance of soil temperature. Our results indicate that mean GSAT may not be the primary driver of treeline in northern Alaska or that its effect is mediated by other more proximate, and possibly non‐climatic, controls. Our model predicts treeline potential in several areas beyond current treelines, pointing to possible routes of treeline advance if unconstrained by non‐climatic factors.

No abstract

No abstract