Modelling habitat suitability for multiple priority weed species to predict invasion hotspots for strategic management in complex landscapes

Established weeds require ongoing, costly management to reduce harm to agricultural and environmental values. Heterogeneous landscapes that encompass substantial variation in land use, vegetation and topographic conditions are often under threat from multiple long-established weeds, whose simultaneous management presents strategic and tactical challenges. Systematic monitoring of weed populations provides more detailed insights into invasion threats than the more readily available presence-only data. We leveraged six years of weed density monitoring and high-resolution environmental data to model habitat suitability for four high-priority weeds across the Australian Capital Territory. We developed habitat suitability maps for African lovegrass, Chilean needlegrass, serrated tussock grass and St John’s wort. Maps were then standardised and combined to derive potential multi-species invasion hotspots and explore patterns of secondary invasion threat. While some species shared environmental predictors such as elevation, each responded to distinct combinations and interactions among environmental predictors. Overall, we found approximately 71% of the study area was suitable for multiple weeds to exceed low densities, but only 11% was sufficiently suitable for two or more species to reach high densities. In those invasion hotspots, habitat suitability was often highest for St John’s wort, and where suitability was highest for either of the other three weeds, St John’s wort typically posed the second highest invasion threat. These findings could improve management efficiency by focussing integrated control of multiple co-occurring species in a small number of priority invasion hotspots. Our study demonstrates how spatially explicit models and analyses based on systematically collected weed density data can inform a multi-species, landscape-scale approach to weed management.