Synthesis reveals that island species–area relationships emerge from processes beyond passive sampling

Leana Gooriah,Alban Sagouis,Jonathan M Chase,Shane A Blowes,Dirk N Karger,Julian Schrader,Holger Kreft,Cang Hui

doi:10.1111/geb.13361

Abstract

AbstractAimThe island species–area relationship (ISAR) quantifies how the number of species increases as the area of an island or island‐like habitat gets larger and is one of the most general patterns in ecology. However, studies that measure the ISAR often confound variation in sampling methodology and analyses, precluding appropriate syntheses of its underlying mechanisms. Most ISAR studies use only presence–absence data at the whole‐island scale, whereas we planned to use a framework that applies individual‐based rarefaction to synthesize whether and how the ISAR differs from the null expectation of the passive sampling hypothesis.LocationFive hundred and five islands from 34 different archipelagos across the world, including oceanic islands, lake islands and forest islands.Major taxa studiedLocal assemblages of plants, invertebrates, herpetofauna, birds and mammals.MethodsWe collated local‐scale species abundance data from multiple archipelagos (median of 12 islands per study) and used a rarefaction‐based approach to synthesize the relationship between island size and (1) sample effort‐controlled rarefied species richness, or (2) an effective number of species derived from the probability of interspecific encounter (an index of community evenness).ResultsWhen we applied rarefaction to control for sampling effort, the numbers of species and their relative abundances across all studies differed from the passive sampling hypothesis. Our measure of evenness also increased with island size, suggesting that the disproportionate effects we observed influenced both rarer and more common species. We found few associations between the slope of this effect and island type or taxon, but we did find that island archipelagos with greater elevational heterogeneity also deviated more from the null expectation than those with less heterogeneity.Main conclusionsUsing a synthetic approach across island archipelagos, we reject the null expectation that passive sampling causes the ISAR and instead suggest that ecological mechanisms leading to disproportionate (non‐random) effects on larger relative to smaller islands are predominant.

Highlights

The island species–area relationship (ISAR) describes how the numbers of species on an island or island-like habitat increases with island area
Given that it is rare for studies to present both information on the total numbers of species per island and data on their abundances from samples within islands, we focused only on datasets with information on the latter; that is, we did not quantify the ISAR of total species richness, which has been well studied and synthesized previously, we expect that it was operating at the scale of the entire island for most of these studies, because this is a very general pattern (e.g., Matthews et al, 2016, 2019; Triantis et al, 2012)
With the combination of our individual rarefaction-based tools for hypothesis testing, in addition to a synthetic approach that allows us to take advantage of the combined information across multiple studies, we were able to reject the null hypothesis that the ISAR largely results only from passive sampling effects; that is, across studies, we found generally more species for a given sample of individuals (Sn), in addition to more even communities, with increasing island area

Summary

| INTRODUCTION

The island species–area relationship (ISAR) describes how the numbers of species on an island or island-like habitat increases with island area. If the null hypothesis of passive sampling is rejected, we can invoke ecological mechanisms that determine the magnitude by which it deviates, and the relative abundances of some species increase (or decrease) more than would have been expected by random sampling Hereafter, we call these “disproportionate effects”, and they can include the influence of area per se (Connor & McCoy, 1979), such as variable colonization–extinction dynamics (e.g., MacArthur & Wilson, 1963, 1967), or population-level processes that tend to be more prominent in communities on smaller rather than larger islands (e.g., Allee effects or demographic stochasticity) (e.g., Courchamp et al, 2008; Lande, 1998). As with Sn above, SPIE can decrease with increasing island area if larger islands support less even communities

| METHODS

Findings

| DISCUSSION