IntroductionClimate change is exerting profound impacts on ecosystems worldwide. In Quebec, Canada, mean annual temperatures are projected to rise by 2°C over the coming decades. In addition, the frequency and intensity of climate extremes are expected to increase, posing significant risks to forest ecosystems and affecting the performance and survival of tree species.MethodsTo assess the potential impacts of climate change on the local adaptation of white spruce, six geographically distant seed sources were evaluated seven years after planting at two test sites with contrasting growing conditions and latitude in Quebec, eastern Canada. Fall phenology was monitored from mid-July to mid-August 2021. For frost tolerance assessment, representative samples were collected in early September. Additionally, two sampling dates at the end of the growing season were used to characterize the dynamics of non-structural carbohydrates (NSC).Results and DiscussionBud set phenology, autumn frost tolerance, and late-season NSC concentrations (fructose, glucose, sucrose, and pinitol) were significantly influenced by test site, seed source, and their interactions. The average bud set initiation occurred one week later at the southern site compared to the northern site, whereas the final stage of bud set was similar between sites. Significant differences among seed sources were observed for certain phenological stages. By early October, all seed sources growing at the southern site showed reduced frost tolerance compared to those at the northern site, with visible needle tissue damage occurring at temperatures below –12°C, highlighting the increased risk of early autumn frost damage at the southern site. Fructose and glucose exhibited contrasting site-specific temporal trends, increasing from September to October at the southern site while decreasing at the northern site. In contrast, sucrose showed the opposite pattern, with the southwestern seed source exhibiting the lowest content, emphasizing the contrasting roles of NSC in sustaining late-season growth under warmer conditions and in regulating dormancy induction and cold hardiness under colder conditions. Overall, our study highlights the crucial role of NSC dynamics in cold tolerance and local adaptation to early autumn frost events. The implications of these findings for assisted migration strategies are discussed.

Pollen analysis is a cornerstone of paleoecological research, yet the minimum number of grains to be counted and identified in a given sediment sample to obtain a statistically robust and relevant vegetation reconstruction remains untested in several contexts. For example, thresholds of 300–500 grains are commonly applied in studies conducted in boreal and northern temperate forests of eastern Canada, but lack empirical validation. We present the first systematic assessment of pollen count requirements from modern samples obtained from ten sites spanning five vegetation zones in eastern Canada, from open boreal woodlands (54°N) to closed-canopy temperate forests (46°N). Assemblage stability, climate reconstruction precision, and morphotype richness were evaluated across incremental counts from 10 to 1000 grains using correlation-based metrics, rarefaction analysis, and two quantitative climate reconstruction methods. We show that minimum pollen counts are context-dependent, determined by vegetation composition and analytical objectives. To reconstruct dominant vegetation assemblages, a total of 250 grains achieve correlation coefficients exceeding 0.95 and capture approximately 70% of the taxonomic richness. In contrast, comprehensive biodiversity assessment and exhaustive detection of rare taxa require substantially higher effort, with 1000 grains capturing 85–95% of total richness. Quantitative temperature reconstructions converge at modest counts, with 420–640 grains needed to achieve 0.5 °C precision. This reflects the strong climatic signal carried by a few dominant taxa (Betula, Pinus, Picea, Alnus), which collectively comprise 80–90% of assemblages. These results challenge the application of a universal count threshold and provide a basis to develop adaptive, objective-driven pollen counting protocols.