A global comparison of plant invasions on oceanic islands

Oceanic Islands have long been considered to be particularly vulnerable to biotic invasis is, and much research has focused on invasive plants on oceanic islands However, findings from ind.aidual islands have rarely been compared between islands within or between biogeographic regions. Ole present in this study the most comprehensive, standardized dataset to date on the global distribut on of invasive plant species in natural areas of oceanic islands. We compiled lists of moderate (5-2:4 cover) and dominant ( > 25% cover) invasive plant species for 30 island groups from four oceanic ref,: oils (Atlantic, Caribbean, Pacific, and Western Indian Ocean). To assess Consistency of plant behaviou- across island groups, we also recorded present but not invasive species in each island group. We tested the importance of different factors discussed in the literature in predicting 'lie number of invasive plant species per island group, including island area and isolaticin, habitat cliversit.3, native species diversity, and human development Further we investigated whether particular invasPi' species are consistently and predictably invasive across island archipelagoi. or -whether island-specific f=.; :tors are more important than species traits in explaining the invasion success of particular species. We found in total 383 non-native spermatophyte plants that were invasive in natural area:, in at least one of the 30 studied island groups, with between 3 and 74 invaders per island group. Of these im.,ar:ers about 50% (181 species) were dominants or co-dominants of a habitat in at least one island group An exn..polation from species accumulation curves across the 30 island groups indicates that the total current flora oi ilviiSive plants on oceanic islands at latitudes between c. 35 N and 35 5 may eventually consist of 500-800;alermatophyte species, with 250-350 of these being dominant invaders in at least one island group The nun er of invaders per island group was well predicted by a combination of human development (rneasurt-a by the gross domestic product (GDP) per capita), habitat diversity.(number of habitat types), island age, and oceanic region (87% of variation explained). Island area, latitude, isolation from continents, number of pres+.:nt, non-native species with a known invasion history, and native species richness were not retained as signiI cant factors in the multivariate models. Among 259 invaders present in at least five island groups, only 9 species were dorninai; invaders in at ' least 50% of island groups where they were present. Most species were invasive only in one 'a a few island groups although they were typically present in many more island groups. Consequently, sink..inty between island groups was low for invader floras but considerably higher for introduced (but nor nea.....arily invasive) Species - especially in pairs of island groups that are spatially close or similar in latitude. He for invasive plants of natural areas, biotic homogenization among oceanic islands may be driven by the recurrent deliberate human introduction of the same species to different islands. while post-introde.cion processes during establishment and spread in natural areas tend to reduce similanty in invader tomtit ititai between oceanic islands. We discuss a number of possible mechanisms, including time lags. pinpaguh. pressure, local biotic and abiotic factors, Invader community assembiy history, and genotypic differences tir t;nay explain the inconsistent performance of particular invasive species in different island groups. 2009 Rtibel Foundation, FM Zurich. Published by Elsevier GmbH. All ghts reserved.