Eva Gril 博士（JSPS外国人特別研究員・北海道大学大学院農学院）

Abstract
In forests, local microclimate temperatures experienced by understory plants differ from macroclimate temperatures measured by weather stations in free-air conditions. Canopies buffer temperature extremes, with lower maxima and higher minima inside dense forests. Using in-situ microclimate measurements in 157 plots across three French forests, we first tested the hypothesis that microclimate can be inferred from plant community composition using ecological indicator values, but this bioindication has a limited applicability for temperature within forests. Secondly, we developed a new statistical approach to model microclimate temperature. Our “slope & equilibrium” method relies on the linear relationship between microclimate and macroclimate temperature. We used forest inventory descriptors to model microclimate buffering (less extreme) or amplification (more extreme) relative to macroclimate fluctuations in 13 deciduous and coniferous forest sites from a long-term monitoring network in France. Thirdly, we used this method to spatialize microclimate temperature at a 10-m resolution over a 2700-ha forest near Blois, using three forest structure metrics derived from airborne LiDAR (light detection and ranging): maximum height, Plant Area Index and Vertical Complexity Index. Finally, we explored the response of understory plants to this microclimate buffering or amplification effect. We observed a shift of microclimate preference along a forest affinity gradient for vascular plants. Bryophytes almost systematically preferred buffered microclimate temperatures. For bryophytes as well as vascular plants, the more thermally buffered, the higher the proportion of forest core specialists. Forest management can enhance or mitigate the effects of climate change on plant communities through microclimate temperature amplification or buffering.