Ecological indicator values of understorey plants perform poorly to infer forest microclimate temperature

Question Ecological indicator values (EIVs) reflect species‘ optimal conditions on an environmental gradient, such as temperature. Averaged over a community, they are used to quantify thermophilization stemming from climate change, i.e. the reshuffling of communities toward more warm‐adapted species. In forests, understorey plant communities do not keep up with global warming and accumulate a climatic debt. Although the causes are still debated, this thermal lag may be partly explained by forest microclimate buffering. For the first time, we test whether community means of EIVs are able to capture microclimate (here, under forest canopies) temperature across, or also within forests. Location 157 forest plots across three French deciduous forests covering a large macroclimatic gradient. Methods To assess whether EIVs can be used to infer the mean and range of microclimate temperature in forests, we measured understorey air temperature for ca. 1 year (10 months) with sensors located 1 m above the ground. We surveyed bryophytes and vascular plants within 400‐m2 plots, and computed floristic temperature from ordinal‐scale EIVs (Ellenberg, Julve) and degree‐scale EIVs (ClimPlant, Bryophytes of Europe Traits) for both temperature and continentality, i.e. temperature annual range. Finally, we fitted linear models to assess whether EIVs could explain the mean and range of microclimate temperature in forests. Results Vascular plant and bryophyte communities successfully reflected differences in mean annual temperatures across forests but largely failed to do so for microclimate variation within forests. Bryophytes did not perform better than vascular plants to infer microclimate conditions. The annual range of microclimate temperatures was poorly associated with ordinal‐scale EIVs for continentality but was positively correlated with degree‐scale EIVs for annual range within lowland forests, especially for vascular plant communities. Conclusion Overall, the capabilities of EIVs to infer microclimate was inconsistent. Refined EIVs for temperature are needed to capture forest microclimates experienced by understorey species. For the first time, we test whether ecological indicator values (EIVs) of forest plant communities are solely able to capture macroclimate or also microclimate temperature. Vascular plant and bryophyte communities successfully inferred macroclimatic differences across forests but largely failed to do so for microclimate variation within forests.