<p>Drivers of global change have direct impacts on the structure of communities and functioning of ecosystems, and interactions between drivers may buffer or exacerbate these direct effects. Interactions among drivers can lead to complex non-linear outcomes for ecosystems, communities and species, but are infrequently quantified. Through a combination of experimental, observational and modelling approaches, I address critical gaps in our understanding of the interactive effects of climate change and plant invasion, using Tongariro National Park (TNP; New Zealand) as a model. TNP is an alpine ecosystem of cultural significance which hosts a unique flora with high rates of endemism. TNP is invaded by the perennial shrub Calluna vulgaris (L.) Hull. My objectives were to: 1) determine whether species-specific phenological shifts have the potential to alter the reproductive capacity of native plants in landscapes affected by invasion; 2) determine whether the effect of invasion intensity on the Species Area Relationship (SAR) of native alpine plant species is influenced by environmental stress; 3) develop a novel modelling framework that would account for density-dependent competitive interactions between native species and C. vulgaris and implement it to determine the combined risk of climate change and plant invasion on the distribution of native plant species; and 4) explore the possible mechanisms leading to a discrepancy in C. vulgaris invasion success on the North and South Islands of New Zealand. I show that species-specific phenological responses to climate warming increase the flowering overlap between a native and an invasive plant. I then show that competition for pollination with the invader decreases the sexual reproduction of the native in some landscapes. I therefore illustrate a previously undescribed interaction between climate warming and plant invasion where the effects of competition for pollination with an invader on the sexual reproduction of the native may be exacerbated by climate warming. Furthermore, I describe a previously unknown pattern of changing invasive plant impact on SAR along an environmental stress gradient. Namely, I demonstrate that interactions between an invasive plant and local native plant species richness become increasingly facilitative along elevational gradients and that the strength of plant interactions is dependent on invader biomass. I then show that the consequences of changing plant interactions at a local scale for the slope of SAR is dependent on the pervasion of the invader. Next, I demonstrate that the inclusion of invasive species density data in distribution models for a native plant leads to greater reductions in predicted native plant distribution and density under future climate change scenarios relative to models based on climate suitability alone. Finally, I find no evidence for large-scale climatic, edaphic, and vegetative limitations to invasion by C. vulgaris on either the North and South Islands of New Zealand. Instead, my results suggest that discrepancies in invasive spread between islands may be driven by human activity: C. vulgaris is associated with the same levels of human disturbance on both islands despite differences in the presence of these conditions between then islands. Altogether, these results show that interactive effects between drivers on biodiversity and ecosystem dynamics are frequently not additive or linear. Therefore, accurate predictions of global change impacts on community structure and ecosystems function require experiments and models which include of interactions among drivers such as climate change and species invasion. These results are pertinent to effective conservation management as most landscapes are concurrently affected by multiple drivers of global environmental change.</p>
Species delimitation and phylogeny of a New Zealand plant species radiation
