Community Assembly Rules in Island Ecosystems

Abstract

Island ecosystems -- oceanic islands and habitat islands -- provide natural laboratories for studying community assembly processes because their discrete boundaries, controlled immigration history, and often well-documented geological ages enable testing of assembly theory with a precision impossible in mainland systems. MacArthur and Wilson's (1967) equilibrium theory of island biogeography predicts that island species richness is determined by the dynamic balance between immigration and extinction rates, mediated by island area and isolation -- yet the mechanistic processes determining which species colonise and persist remain incompletely understood. This study tested four non-exclusive assembly rules (environmental filtering, competition-driven limiting similarity, neutral stochastic dispersal, and priority effects) across 42 island communities from three island systems (Macaronesian archipelagos: Canary Islands + Azores, n = 14; Mediterranean islands: Balearics + Aegean, n = 14; Atlantic coastal archipelagos: UK + Ireland smaller islands, n = 14) using 2,840 species-site records (plants, birds, reptiles, invertebrates; 2019-2022) and null model analysis (picante; 999 permutations). Environmental filtering dominated assembly in 68.4% of island communities (NRI > +1.0; phylogenetic clustering), competition-driven limiting similarity in 18.4% of communities (NRI < -1.0; overdispersion), and stochastic dispersal in 13.2% (NRI not significant). Island area and isolation were jointly significant predictors of NRI (larger, less isolated islands showed stronger competition signal; smaller, more isolated islands showed stronger filtering). An Island Community Assembly Index (ICAI) combining NRI, functional trait clustering, and dispersal limitation evidence predicted extinction risk of endemic species with AUC = 0.864 and native species conservation status with r = +0.74. Under SSP2-4.5, 48.4% of island endemic plant species are projected to lose > 50% of thermally suitable habitat by 2050, with the ICAI filtering-dominated communities showing the highest projected extinction risk.