Pervasive human-mediated large-scale invasion: analysis of spread patterns and their underlying mechanisms in 17 of China's worst invasive plants

Biological invasions constitute a major component of human-induced environmental change and have become a world-wide problem threatening global biodiversity and incurring massive economic costs. Consequently, research on biological invasions proliferates, placing a major emphasis on species traits and habitat characteristics associated with successful invasion. Yet, the mechanisms underlying rapid spread and the resulting patterns remain largely unexplored. Using data collected since 1980 and earlier at the county level all over China, we studied the contribution of potential dispersal vectors – railroads, rail stations, roads, general human activity, rivers and winds – to the spread of 17 of China's worst invasive plant species. Focusing on long-distance dispersal events, we calculated the minimal arrival speed for the first record of each species in each county. We also developed and applied a new method to account for observation bias due to the proximity to roads, using observational data of 776 native (non-invasive) plant species throughout China. We found that human-related vectors are accountable for the vast spread of all 17 invasive plant species we examined. Spread patterns were characterized by long jumps of tens to hundreds of kilometres and extremely fast average spread rates of roughly 2–4 km per year, and a very broad range (0·1–128·2 km per year) with high variability between years. These rates are much higher than those expected from classic dispersal vectors such as water, wind or animals. Commonly used fat-tailed dispersal kernels did not fit the observed distribution of long jumps for any species. Synthesis. We found pervasive empirical evidence for the overriding role of humans in the large-scale spread of invasive plants from multiple taxa. The observed spread patterns differ significantly from those portrayed in the literature, emphasizing the need to develop new frameworks to explore large-scale spread in general and invasive spread in particular. With public data sets of invasive species observations becoming increasingly more available, the time is ripe to go beyond exploration of species traits and habitat suitability and to examine the actual patterns and the mechanisms of large-scale invasive spread, even at a scale of thousands of kilometres over land.