Predicting spatially heterogeneous invasive spread: Pyracantha angustifolia invading a dry Andean valley in northern Argentina.
Understanding the drivers of invasive species spread is key to designing optimal management programmes for controlling them. Population models, parameterized from demographic and dispersal data, are useful for simulating invasion when long-term observations are lacking, which is particularly important when invasions are occurring rapidly. We aimed to understand how the spread of Pyracantha angustifolia (an invasive ornamental shrub) is influenced by heterogeneity in local conditions within a dry inter-montane valley in northwestern Argentina that is currently experiencing rapid urban expansion and concomitant reduction in cattle farming. Field surveys demonstrated that individuals growing in shrublands have a higher fecundity than those in grassland and rocky habitats. Furthermore, grazing pressure of cattle and horses substantially reduces fecundity and impacts the relationship between age and fecundity. We incorporated our field-estimated habitat and age-specific fecundities into both analytical integro-difference equation models and individual-based models to predict rates of spread across each habitat type in the presence or absence of grazing pressure. Results indicate that the rate of spread would be substantially higher in shrublands (by up to 33%) and highlight that grazing pressure can, through direct consumption of seeds by livestock, substantially depress the rate of spread (by up to 53%). These results suggest that shrubland areas in the valley are most vulnerable to invasion and that grazing may help to reduce the impact of Pyracantha. This suggests that a synergism between urban expansion and reduction in domestic livestock density contributes to the spread of an invasive species in this environment.