Functional traits fail to predict long-term responses to climate change

Aim. Environmental conditions strongly affect the distribution and abundance of species via complex forces. Shifts in environmental conditions and differences in the speed and scale of these effects complicate our efforts to infer how species will respond to future environmental change. We test how 18 functional traits affect plant species responses to gradients in environmental conditions and 50-year shifts in climate. Location: We analyzed 50-year shifts in the distribution and abundance of 153 plant species distributed across 284 sites in Wisconsin, USA. Time period. 1950s to 2000s. Major taxa studied. Vascular plants (much of the flora of NE North America). Methods. We used random forest and integrated hierarchical mixed models to test how plant abundances (and 50-year changes in abundance) track gradients in overstory, soil, and climatic conditions. Results. Within study periods, plant abundances reflect gradients in environmental conditions. Leaf traits affected local abundance (both directly and via trait-environment interactions) in the 1950s and 2000s. Strong soil and temperature effects in the 1950s have weakened while precipitation effects have strengthened. Although we expected these models to also predict how plants would respond to shifts in climate, they did not. Conclusions. Lags in species responses, increases in the stochastic forces affecting community assembly, and other forces limit the ability of models fitted to static data (e.g., space-for-time substitutions) to predict how plant species will respond to long-term shifts in environmental conditions. We must therefore be cautious about applying trait-based species distribution models to predict how climate change will affect species distributions and community structure. ### Competing Interest Statement The authors have declared no competing interest.