Consideration of dispersal processes and northern refugia can improve our understanding of past plant migration rates in North America

• Published in: Journal of biogeography 2015-09, Vol.42 (9), p.1677-1688
• Main Authors: Snell, Rebecca S., Cowling, Sharon A.
• Format: Article
• Language: English
• Subjects: Dynamic vegetation model; glacial refugia; long-distance dispersal; LPJ-GUESS; migration; Migration and dispersal processes; North America; post-glacial plant colonization
• ISSN: 0305-0270; 1365-2699
• DOI: 10.1111/jbi.12544

Aim: According to the palynological record, rapid plant migrations following the retreat of the last glacier were common in North America and Europe. However, providing an explanation for these rapid migration rates has been challenging considering modern-day seed dispersal distances. We used the newly developed seed dispersal functionality in a hybrid dynamic global vegetation model (DGVM) to simulate two theories that have been proposed to explain rapid plant migrations: long-distance seed dispersal and northern refugiai populations. Location: Idealized landscapes representing temperate and boreal regions of North America. Methods: Vegetation migration rates for three species (Acer rubrum, Vagus grandifolia and Picea glauca) were simulated in response to climate change across a landscape. All simulations included long-distance seed dispersal; however, we compared landscape colonization rates both with and without the presence of northern refugiai populations. Results: For all three species, the colonization rates were faster when there was a northern réfugiai population. Increasing the number of locations of refugia further increased the rate of landscape colonization, and this was most effective when réfugiai populations were spatially separated. The perceived migration rates (i.e. the time it took to spread the furthest distance away from the southern refugium) were approximately twice as fast when a refugium was present. For example, A rubrum had a perceived migration rate of 119 m yr⁻¹ without réfugia and a perceived migration rate of 204 m yr⁻¹ with a northern refugium. Main conclusions: The simulated migration rates that matched the rapid migration rates observed in the pollen record for Acer were only achieved when both long-distance dispersal and northern réfugiai populations were included in the simulations. Even with réfugiai populations, migration rates for P. glauca and F. grandifolia were slower than historical rates. The results of this study are consistent with the hypothesis that plant migration rates calculated from the pollen record overestimate the ability of most species to track rapid climate change.