While long-distance dispersal is thought to drive invasion speeds across a landscape, once present in a region, the natural infilling process can be quite slow. We investigated the regional spread and infilling of Microstegium vimineum, an invasive grass species, in central Pennsylvania. Prior studies had demonstrated that local unaided dispersal is extremely slow, but road maintenance has the potential to move seeds several hundred meters. Extreme long distance dispersal events are likely driven by rare movement of seeds adhering to vehicles. We developed a spatially explicit landscape model of rare long-distance dispersal via vehicle, intermediate scale spread via road maintenance and local unaided spread of M. vimineum, to investigate the relative contributions of different scales of dispersal. Data from field experiments were used to parameterize the models. We contrasted models including local dispersal only, local and intermediate dispersal, all three scales of dispersal and just local and long-distance dispersal. We examined the resulting population sizes, overall extent of the invasion and the extent to which the road network was invaded. We also examined whether proposed management techniques aimed at reducing local reproduction or reducing intermediate spread could effectively reduce the invasion or protect ecologically valuable areas of the forest.
Results/Conclusions
As expected, local dispersal only produced the smallest invasion in terms of the three metrics examined, and the models with all modes of dispersal had the largest invasions (the average extent of invasion was 17 times larger than local dispersal only). What was more surprising was that the model lacking only intermediate spread led to relatively modest invasions; the average extent of the invasions was only 11 % compared to the invasion when using of all three modes of dispersal. It appears that intermediate scale dispersal is responsible for the commonly observed infestation of the entire road network; previous research has documented that road grading is a likely mechanism of this scale of spread. In terms of management efforts, no treatments were effective at reducing overall spread, but some treatments (grading less often or suppressing reproduction in buffer areas) successfully protected sensitive areas. Road creation and maintenance are necessary parts of our current society that we expect will continue, and these results highlight the need for careful consideration of the increase in plant invasion associated with roads.