Migration‐based simulations for Canadian trees show limited tracking of suitable climate under climate change

Abstract

Abstract Aim Species distribution models typically project climatically suitable habitat for trees in eastern North America to shift hundreds of kilometres this century. We simulated potential migration, accounting for various traits that affect species' ability to track climatically suitable habitat. Location Eastern Canada, covering ~3.7 million km 2 . Methods We simulated migration‐constrained range shifts through 2100 using a hybrid approach combining projections of climatically suitable habitat based on two Representative Concentration Pathways (RCP4.5, RCP8.5) for three time periods and two species distribution modelling approaches with process‐based models parameterized using data related to dispersal ability and generation time. We developed a unique “migration kernel” that uses seed dispersal traits and observed migration velocities to obtain kernel shape and dispersal probabilities for each tree species. We then calculated lags between the migration‐constrained range limits obtained through simulations and limits of climatically suitable habitat. Results All species demonstrated northward range shifts at the leading edge of their simulated distribution through 2100, but the magnitude and rate of that shift varied by species and time period. Climatically suitable habitat limits were found to be north of simulated distribution limits across both RCPs, with lags increasing through time. On average, simulated distribution that remained within climatically suitable habitat declined more under RCP8.5 than RCP4.5, with large areas of the rear edge of the simulated distribution becoming partially or completely climatically unsuitable for many species. Main conclusions Climatically suitable habitat limits projected for 2100 far exceeded migration‐constrained range limits for all 10 tree species, particularly for temperate species. This study underlines the limited extent to which tree species will track climate change via natural migration. Integrating observed migration velocities, seed dispersal and generation time with SDM outputs allows for more realistic evaluations of tree migration ability under climate change and may help orient forest conservation and restoration efforts.