Biological invasions have become one of the main drivers of habitat degradation and a leading cause of biodiversity loss in island ecosystems worldwide. The spread of invasive species poses a particular environmental threat on the islands of the Mediterranean Basin, which are hotspots of biodiversity and contain rare habitats and endemic species, especially on small islands, which are highly vulnerable to biodiversity loss. \ud Following a recent survey, in this paper we aim to provide an overview of the present-day non-native vascular flora of small Mediterranean islands based on a sample of 37 islands located in the middle of the Mediterranean sea, off the coast of Italy. By comparing the current data with those gathered during a previous survey conducted in the same study area, we also aim to highlight the main changes that have occurred in non-native plant species diversity, establishment and distribution in recent years, and to present a first general overview of the most prominent plant taxa in the island’s introduced flora, focusing on those most responsible for these changes and those that pose the greatest environmental threats. \ud We recorded 203 non-native plant species, 147 of which have established on at least one of the islands investigated. Overall, we detected a sharp increase in the number of species, in their levels of establishment and in the extent of their distribution within the study area in recent years. This may be explained by the intensification of research on plant invasions, as well as to new introduction, escape, establishment and invasion events on the islands in recent decades. The most remarkable plants detected include acacias and succulents, two groups that appear to be emerging very rapidly and to be posing new threats to the conservation of the islands’ natural environment, especially the genus Carpobrotus, whose spread into natural habitats containing rare and endemic taxa is seriously threatening biodiversity on both a local and global scale.\ud On the whole, our results show that the plant invasion phenomenon in the study area has in recent years intensified considerably. As this process seems likely to continue, we should expect more establishment events in the future and the further spread of species that are already present. This is of particular conservation concern on the islands investigated in this survey, which are rich in endemisms but have been facing deep socio-economic and environmental transformations in these last decades as a consequence of the abandonment of traditional management practices and the development of tourism. Our study thus confirms that plant invasions on Mediterranean islands are a serious environmental problem that threatens biodiversity conservation not only in the Mediterranean biogeographic region but also on the global scale, and highlights the need to further increase efforts aimed at preventing, controlling or mitigating the effects of plant invasions in island ecosystems

Invader success and ecosystem impact are both expected to be largely driven by the functional trait distinctiveness of the resident species relative to the invaded communities. To understand the importance of trait distinctiveness for plant invasions, and the native community's trait response to the invasion, it is key to measure multiple traits simultaneously, and to incorporate intraspecific trait variation. Here we explored multidimensional patterns of inter‐ and intraspecific trait variation during the invasion of two functionally contrasting species. We constructed multidimensional trait hypervolumes for the invaders, their invaded communities, and uninvaded reference communities. The functional distinctiveness hypothesis predicts that invaders will occupy a mostly unique part of the trait hypervolume and that invasion will shift the trait composition of the native community to minimize trait overlap with the invader. Impatiens glandulifera was characterized by acquisitive traits, while Rosa rugosa had more resource conservative traits, reflecting their respective invaded habitats. Both invaders showed relatively little hypervolume overlap with the uninvaded communities, although this overlap was higher for R. rugosa (31.9%) than for I. glandulifera (14.3%). Both invaders affected community traits, mainly due to intraspecific trait changes of the resident species. As expected, invasion by R. rugosa reduced trait overlap between invader and community to 18.3%. Invasion by I. glandulifera, however, resulted in an increased trait overlap to 26.7%. In both cases the community traits shifted towards a more resource acquisitive strategy following invasion, irrespective of the invader's trait composition. This suggests that invasion‐driven community‐level intraspecific trait shifts are likely caused by increased competition for light under the invader, rather than by niche overlap. These, at first sight contradictory, results demonstrate the need for better trait‐based invasion and community ecology theory. Our results furthermore show that invaders can shift the trait occupancy of resident communities.

No abstract

Functional traits can help elucidate and predict the impact of invasive plant species on ecosystem functioning. Yet, this approach requires comprehensive and labour‐intensive trait collection campaigns, covering intraspecific trait variation of both the invader and native species in the invaded community. One potential way to overcome these logistic constraints is using hyperspectral remote sensing technology to efficiently quantify functional trait values. Although such spectrally derived or ‘optical’ traits are known to closely link to directly measured functional traits, little research has explored how well these optical traits perform in assessing invader‐induced ecosystem impact. Here, we explored the trait‐mediated impact of the invasive Rosa rugosa on litter decomposition and evaluated whether optical traits perform equally well as directly measured traits in predicting litter decomposition variation. We collected data on species‐specific functional traits, leaf hyperspectral reflectance and standardized ‘tea bag index’ litter decomposition across 25 invaded and 25 uninvaded coastal grassland plots. The selected traits were all potentially related to litter decomposition and covered the leaf economics spectrum, additional leaf structural components and competitive ability. Optical traits were quantified through a combination of a physical radiative transfer model inversion and vegetation indices calculations. Invasion significantly increased the stabilization factor, i.e. the amount of resulting recalcitrant litter. Invader impact on litter decomposition could be entirely explained by changes it induced in the functional traits of the native community, rather than by the invader's traits itself. More specifically, the invader pushed the invaded community towards traits associated with high litter quality. Optical traits performed equally well as directly measured traits in explaining the invasion impact on the stabilization factor (R2 = 41.9% vs. 38.5%). Furthermore, the interpretation of the results based on optical traits resulted in a similar functional understanding of the invader impact. Synthesis. Our results indicate the potential of hyperspectral data to explain changes in ecosystem functioning. The combination of radiative transfer models and vegetation indices allowed to extract all relevant trait information from the hyperspectral data. This framework thus presents a practical shortcut to assess relevant leaf traits, requiring only a limited amount of field trait measurements.

Plant monocultures growing for extended periods face severe losses of productivity. This phenomenon, known as ‘yield decline’, is often caused by the accumulation of above- and belowground plant antagonists. The effectiveness of plant defences against antagonists might help explaining differences in yield decline among species. Using a trait-based approach, we studied the role of 20 physical and chemical defence traits of leaves and fine roots on yield decline of 18-year old monocultures of 27 grassland species. We hypothesized that yield decline is lower for species with high defences, that root defences are better predictors of yield decline than leaf defences, and that in roots, physical defences better predict yield decline than chemical defences, while the reverse is true for leaves. We additionally hypothesized that species increasing the expression of defence traits after long-term monoculture growth would suffer less yield decline. We summarized leaf and fine root defence traits using principal component analysis and analysed the relationship between defence traits mean as a measure of defence strenght and defence traits temporal changes of the most informative components and monoculture yield decline. The only significant predictors of yield decline were the mean and temporal changes of the component related to specific root length and root diameter (e.g. the so called collaboration gradient of the root economics space). The principal component analysis of the remaining traits showed strong trade-offs between defences suggesting that different plant species deploy a variety of strategies to defend themselves. This diversity of strategies could preclude the detection of a generalized correlation between the strength and temporal changes of defence gradients and yield decline. Our results show that yield decline is strongly linked to belowground processes particularly to root traits. Further studies are needed to understand the mechanism driving the effect of the collaboration gradient on yield decline.