COS 94-3 - Canopy/understory phenological mismatch differs between continents in response to climate warming

Background/Question/Methods

Understory forest plants are strongly limited by access to light throughout the growing season, and many species in deciduous forests have evolved strategies to make use of ephemeral, seasonally-dynamic light availability. However, anthropogenic climate change is fundamentally changing the forest light regimes to which native species have adapted: community composition is changing; invasive non-native species exhibit novel phenological strategies, frequently leafing earlier in spring than native species; and climate change induces species-specific changes in the timing of overstory canopy leaf out. Each of these changes alone or in combination could cause phenological disruption or mismatch – uneven changes to the relative timing of life events among interacting species – but the ecological consequences of these disruptions are only partially understood. Furthermore, the relatively few studies investigating understory/overstory mismatch have been limited by spatial extent. Here, we scored wildflower and canopy tree herbarium specimens from temperate forests in North America, Europe, and east Asia in an effort to identify if phenological sensitivity to spring temperature differs between the two groups and if these differences vary between continents. We also investigated whether and how the inclusion of spatial autocorrelation in phenology models affect our results and conclusions.

Results/Conclusions

Understory wildflowers and canopy tree species were both significantly sensitive to spring temperature, with sensitivities ranging from -2.3 to -3.5 days/°C. Phenological sensitivity did not significantly differ between wildflowers and trees on any continent in the best-fit models, contrary to previous studies which have documented relatively greater sensitivity for trees (and thus projections of reduced access to light availability under climate change conditions). Instead, we found almost equal sensitivity in North America and relatively greater sensitivity for wildflowers compared to trees in Europe and Asia. This suggests that wildflowers on the latter two continents will instead gain access to spring light under warming springs.

We found that spatial autocorrelation significantly improved model fit across the board and found vast improvements in the correlation between predictions and observations when included. However, it considerably changed the direction and magnitude of our predictions. For example, without spatial autocorrelation, our models estimate that North American trees are significantly more sensitive to spring temperature compared to wildflowers, consistent with previous studies. Our results suggest that 1) spatial autocorrelation is important for herbarium-based phenological models, and 2) overstory/understory mismatch differs across continents, suggesting that climate change will affect access to understory light in different ways.