Testing multiple assembly rule models in avian communities on islands of an inundated lake, Zhejiang Province, China

Aim A fundamental question in community ecology is whether general assembly rules determine the structure of natural communities. Although many types of assembly rules have been described, including Diamond’s assembly rules, constant body-size ratios, favoured states, and nestedness, few studies have tested multiple assembly rule models simultaneously. Therefore, little is known about the relative importance of potential underlying factors such as interspecific competition, inter-guild competition, selective extinction and habitat nestedness in structuring community composition. Here, we test the above four assembly rule models and examine the causal basis for the observed patterns using bird data collected on islands of an inundated lake. Location Thousand Island Lake, China. Methods We collected data on presence–absence matrices, body size and functional groups for bird assemblages on 42 islands from 2007 to 2009. To test the above four assembly rule models, we used null model analyses to compare observed species co-occurrence patterns, body-size distributions and functional group distributions with randomly generated assemblages. To ensure that the results were not biased by the inclusion of species with extremely different ecologies, we conducted separate analyses for the entire assemblage and for various subset matrices classified according to foraging guilds. Results The bird assemblages did not support predictions by several competitively structured assembly rule models, including Diamond’s assembly rules, constant body-size ratios, and favoured states. In contrast, bird assemblages were highly significantly nested and were apparently shaped by extinction processes mediated through area effects and habitat nestedness. The nestedness of bird assemblages was not a result of passive sampling or selective colonization. These results were very consistent, regardless of whether the entire assemblage or the subset matrices were analysed. Main conclusions Our results suggest that bird assemblages were shaped by extinction processes mediated through area effects and habitat nestedness, rather than by interspecific or inter-guild competition. From a conservation point of view, our results indicate that we should protect both the largest islands with the most species-rich communities and habitat-rich islands in order to maximize the number of species preserved.