Habitat filtering and dispersal ability determine across-scale community turnover and rarity in Macaronesian island spider communities

Aim: Much research has focused on the separate and combined effects of habitat and geography on species communities, both in mainland and island ecosystems. However, few studies have looked into the differences among communities inhabiting different island habitats. Here we aim to determine the contributions of habitat filtering and dispersal to the differences in taxonomic structure of island communities across geographic scales. Location: Sixty plots in two habitats (forest and dry habitat) across eight islands of the Macaronesian archipelagos of the Azores, Madeira, Canary Islands and Cabo Verde. Taxon: Spiders (Araneae). Methods: We generated community data by using the optimised and standardised COBRA sampling protocols. We tested the differences in three beta diversity metrics ( βTotal, βReplacement and βRichness) for each habitat and dispersal category separately, across geographic scales through nested non-parametric PerMANOVA. We tested if dispersal and habitat influenced differences between the communities of the same area by applying a PERMDISP to βreplacement, and using the distances in linear mixed models. We tested the effects of habitat and dispersal on species relative abundances (SADs) and rarity by building Gambin models for each community and dispersal group separately. Results: Communities grouped according to archipelago and habitat, in terms of taxonomic similarity. In general, β diversity increased with geographic scale, and was greater in dry habitats. βReplacement among communities of the same region was greater in dry habitats than in forests, and this difference was stronger for rare ballooners. Dry habitat communities showed more species with low abundances (rare species) at different spatial scales. Main conclusions: Our findings reveal that habitat type does not only condition the processes behind community assembly but also the scale at which they occur. Indeed, our findings are highly relevant for theories on across and within island community assembly as well as for biodiversity conservation.Field work was supported by two projects: 1) the ERA-Net Net-Biome research framework, Canary Islands Government ACIISI grants SE-12/02 (PO), SE-12/03 (JCC), SE-12/04 (BE), co-financed by FEDER, Portuguese FCT-NETBIOME grant 0003/2011 (PB); and 2) FCT MACDIV – FCT PTDC/BIABIC/0054/2014. Open Access was financed by the project UID/BIA/00329/2020. The National Parks of Terceira and Pico (Azores), the Instituto das Florestas e da Conservação da Natureza (Madeira), the Cabildos of Tenerife and La Gomera, Garajonay National Park (La Gomera) as well as the Cabo Verde Government provided permits for specimen collection in protected areas. IRA was funded by national funds through FCT – Fundação para a Ciência e a Tecnologia, I.P., under the Norma Transitória DL57/2016/CP1375/CT0003.Peer reviewe