Abstract
The rate at which invading organisms disperse into novel habitats is fundamental to their distribution and abundance. Forecasts of spread often assume that invasion speed is constant through time and among directions but, depending on the extent to which this assumption is violated, the efficacy of delimitation surveys and eradication programs could suffer. Knowledge of the mechanisms underlying spatiotemporal variation in spread could help refine forecasts and guide management, particularly in the early stages of invasions. We investigated rates of spread by laurel wilt, one of the most damaging non-native forest pests in North America, using three standard approaches (effective range radius, distance regression, and boundary displacement) and evaluated the strength and drivers of variation in directional spread (i.e., anisotropy). Estimates of mean annual spread varied from 24 to 40 km/yr, but spread was highly anisotropic with invasion speeds reaching approximately 100 km/yr south, 80 km/yr west, and 50 km/yr north, a pattern that we attribute to the abundance of host redbay trees and warmer temperatures fostering rapid southern and western spread. This pattern-quicker spread of laurel wilt from the point of introduction into areas forecasted as highly suitable for its persistence-suggests that establishment location might have a major influence on rates of anisotropy. Our findings underscore the utility of habitat suitability modeling-in which host availability and suitable climate are widely used to forecast establishment risk-for identifying areas into which spread will proceed most rapidly following establishment of a new invader and/or a satellite population via a long-distance dispersal event. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s10530-023-03069-5.