Abstract: Climate change has altered the global temperature regimes leading to warmer temperatures occurring earlier in spring in many temperate regions. This has induced an earlier budbreak making trees susceptible to late spring-frost events, however, information on species-specific late spring-frost tolerance is only available from observational studies. Here, we implemented a quantitative study on late spring-frost tolerance determined by in vitro leaf-level measurements of three temperate broad-leaved tree species via the assessment of the maximum quantum yield efficiency of the photosystem II (F_v/F_m). We investigated to what extent in vitro measurements conducted one day before a late spring-frost event can predict the in vivo damage caused by a cold snap. Fraxinus excelsior showed the lowest in vitro tested tolerance to late spring-frost, and the leaves lost 50% of F_v/F_m (LT50) at + 0.60 ± 0.26 °C. The damage induced by the cold snap the following day (minimum temperature of − 3.28 °C) was a fatal decline of F_v/F_m to 5.8% of the maximum. The other two species, namely Fagus sylvatica and Quercus robur, were characterized by LT50 of − 0.17 ± 9.99 °C and − 2.29 ± 1.11 °C, respectively. The cold snap induced less damage, F_v/F_m values declined to 46.9% and 53.5% of the maximum in the two species, respectively. The in vitro measurements precisely predicted the damage caused by the late spring-frost event. We suggest that in vitro estimated LT50 values can be used as a comparative leaf trait as it has high predictive power for tree species performance after late spring-frost.

Key words: Cold snap, Spring phenology, Leaf unfolding, Frost damage, Temperate trees