• climate gradients;
  • fragmented landscapes;
  • grazing pastures;
  • invasive species;
  • land use;
  • landscape hierarchy;
  • mixed-effect models


  1. To date, there has been a focus on biological aspects of invasion rather than landscape-scale processes within human-modified landscapes. Applying a multi-scale approach may improve predictive models of invasion in fragmented landscapes as well as management.
  2. We tested a priori models of the relative importance of landscape (10 000 to 30 000 ha), site-scale (<0·3 ha) factors and climate gradients, and their interactions, on the cover of exotic grasses in forest fragments. We tested the models for two exotic pasture grasses Cenchrus ciliaris and Panicum maximum in fragmented agricultural landscapes of subtropical eastern Australia. Mixed-effect models were applied to quantify the multi-scale effects of land use, forest area and fragmentation, vegetation structure, soil and climate attributes and their interactions on the ground cover of the two exotic grasses.
  3. We used model averaging to account for model and parameter uncertainty and tested for spatial autocorrelation and the model fit in model residuals.
  4. There was strong evidence that the area of exotic pasture grass in the landscape and climate gradients had a strong influence on cover observed at the site scale, but the climate influence was species specific. Cenchrus ciliaris cover was higher in forest sites within landscapes converted to exotic pastures and with high mean minimum temperatures and sites with low canopy cover. In contrast, P. maximum cover was primarily influenced by high mean annual rainfall, with moderate positive effects from area of exotic pasture in the landscape and high minimum temperatures.
  5. Synthesis and applications. This study highlights several key issues that need to be considered when prioritizing management of invasive species in fragmented landscapes. First, the propagule supply in the landscape and climate gradients are key factors influencing the spread of exotic grasses into forest fragments. Second, managing canopy cover is critical to enhancing C. ciliaris invasion resistance, but not for P. maximum. We conclude that a multi-scale approach provides the conceptual and applied basis for an improved understanding and management of plant invasions in fragmented landscapes.