|dc.description.abstract||Biological invasions have noticeably accelerated with increases in global trade, and international travel. While many non-native species cause changes in both managed and natural ecosystems, and pose a major threat to native biodiversity and economic activity worldwide. To prevent the establishment and economic and/or environmental impact of invasive organisms, and protect productive and natural ecosystems around the world, risk assessment is necessary. To accomplish that, knowledge about recipient ecosystem characteristics and the potential invaders that influence invasion success needs to be increased. Unfortunately, because economic pests usually attract more scientific attention there is a general lack of research on non-native arthropod impacts on natural systems.
The main aim of this research was to develop methods to improve risk assessment of the potential establishment and impact of non-native arthropod species that threaten natural ecosystems. The objectives were to, 1) compile records of non-native beetles and spiders established in natural ecosystems in North America, a continental ecosystem, and New Zealand, an oceanic island ecosystem, to compare their impacts on those ecosystems, 2) determine whether floristic similarity between donor and recipient regions and floristic diversity, are useful complements to climatic similarity for predicting potential establishment, 3) determine if plant host phylogenetic affinities between donor and recipient regions can predict the relative impacts on individual plant species and ecological systems, and, 4) review and test different impact classification frameworks using structured expert elicitation methods.
The comparison of the level of impact of non-native beetles and spiders in natural ecosystems of North America and New Zealand showed that North American natural ecosystems seem more vulnerable to phytophagous insect invasions than those in New Zealand. New Zealand natural ecosystems, however, appear more sensitive to invertebrate generalist predators, such as spiders. While these results seem to indicate that New Zealand’s native ecosystems might be more invasible by host-generalist invertebrates than continental North America. The relationship between biogeographic origin of the non-native species and their level of impact in each ecosystem, was also studied. But, unexpectedly, there was no clear relationship between the two variables in either ecosystem studied.
Other factors proposed to improve risk assessment of the establishment and impact of non-indigenous invertebrate herbivores, were investigated. First, the influence of floristic similarity, floristic diversity and climatic similarity on non-native beetle establishment success was studied. Different models showed that including floristic and climate variables in the same model better predicted non-native beetle presence than models that included the variables separately. However, for the species in this study, floristic, similarity between regions in combination with floristic diversity within a region better explained the presence of non-native beetle species than climate similarity alone.
Plant phylogenetic affinities between donor and recipient regions is an important indicator of the establishment success of non-native herbivore species in the recipient region. As expected, the host-range of the non-native herbivore species established in New Zealand are phylogenetically constrained. However, the probability of feeding on new host plants, and therefore the phytosanitary risk associated with the species, varied among classes of herbivore taxa and among the orders of phytophagous insects. In addition, phylogenetic affinities analysis between potential donor and recipient regions was shown to have predictive power that may allow the identification of potential herbivore pests of native ecosystems as well as which plant genera might be susceptible.
Finally, the analyses in this research raised questions about how best to assess the impact level and therefore the level of risk in the invaded area. A structured expert elicitation process was used in this to assess impact and was shown to increase agreement between the experts using the selected frameworks. However, this research clearly showed that developing a generic quantitative framework should be the subject of urgent research.
Overall, this research has contributed to increasing knowledge and awareness of the range of factors that can determine or influence non-native invertebrate arthropod species establishment success and subsequent impact. Particularly floristic similarity and plant phylogenetic affinities analysis that should be useful additional variables for risk assessors in biosecurity regulatory authorities around the world, to rapidly evaluate the risk of establishment of newly detected invaders.||en