Abstract
To answer the long-standing question if we can predict plant invader success based on characteristics of the environment (invasibility) or the invasive species (invasiveness), or the combination of both, there is a need for detailed observational studies in which habitat properties, non-native plant traits, and the resulting invader success are locally measured. In this study, we assess the interaction of gradients in the environmental and trait space on non-native species fitness, expressed as seed production, for a set of 10 invasive and noninvasive non-native species along a wide range of invaded sites in Flanders. In our multidimensional approach, most of the single environmental gradients (temperature, light availability, native plant species diversity, and soil fertility) and sets of non-native plant traits (plant size, photosynthesis, and foliar chemical attributes) related positively with invader seed production. Yet correlation with seed production was much stronger when several environmental gradients were assessed in interaction, and even more so when we combined plant traits and habitat properties. The latter increased explanatory power of the models on average by 25% for invasive and by 7% for noninvasive species. Additionally, we report a 70-fold higher seed production in invasive than in noninvasive species and fundamentally different correlations of seed production with plant traits and habitat properties in noninvasive versus invasive species. We conclude that locally measured traits and properties deserve much more attention than they currently get in invasion literature and thus encourage further studies combining this level of detail with the generality of a multiregion and multispecies approach across different stages of invasion.