Requirements of plant species are linked to area and determine species pool and richness on small islands

Abstract

Questions:Small islands are outstanding model systems to study community assembly. Due to harsher environmental conditions on smaller islands compared to larger ones, environmental filtering may preclude some species, potentially resulting in island size-dependent species pools. We tested whether the species pool size follows a similar species-area relationship as the observed richness. This can provide new insight into community assembly processes and the elusive small-island effect (SIE), which states that species richness on smaller islands is less dependent on area than on larger islands.Location: Raja Ampat Archipelago, Indonesia. Methods:We studied the woody vegetation on sixty small islands ranging from 3 m 2 to 11,806 m 2 . For each recorded species, we estimated its area requirements and compared them against random colonization models. We developed a novel method to calculate probabilistic species pools for each island. We compared different species-area models for observed species richness and our index of species pool size to test whether the SIE results from differences in species pool size. Results:We found that most species were restricted to islands significantly larger than expected from random colonization. The occurrence probability of all species increased with island size, indicating a lack of species that are specialized to the conditions on small islands. We found a SIE in observed species richness, but not in species pool size. Conclusion:Woody plants in the studied island system have specific requirements that are linked to island area and determine island-specific species pools. Lower community completeness on smaller islands compared to larger ones indicated that the SIE is shaped by local limiting processes that have no impact on the species pool, but control how much of it is realized on an island. Together, these results clearly indicate non-random plant community assembly on small islands. K E Y W O R D S community completeness, dark diversity, null model, occurrence probability, small-island effect, species pool, species richness, species-area relationship 600 | Additional supporting information may be found online in the Supporting Information section at the end of the article. Appendix S1. Species-area relationships of observed and estimated species richness for the studied islands Appendix S2. Relative importance of six explanatory variables explaining observed species richness on the studied islands Appendix S3. Island characteristics of the studied islands Appendix S4. Empirical and simulated species occurrence probabilities for all species on the studied islands Appendix S5. Model support for three species-area relationship models for the species pool size and the observed species richness How to cite this article: Schrader J, König C, Moeljono S, Pärtel M, Kreft H. Requirements of plant species are linked to area and determine species pool and richness on small islands. J Veg Sci. 2019;30:599-609. https ://doi.

Plants on small islands revisited: the effects of spatial scale and habitat quality on the species–area relationship

“…3b-e). The presence of the SIE was further supported by a null model that indicated that small islands have less species present than expected under random colonisations (Supplementary material Appendix 10; see also Schrader et al 2019). For α 2 , the Lomolino model performed best (note that five other models were similarly well supported with ΔAICc <2), but was less well supported than for γ 1 and γ 2 (Fig.…”

“…The high variability of species richness at local scales implies that the false detection (if by chance more islands with lower species richness are sampled), or non-detection (if more islands with higher richness are sampled) of the SIE could easily occur if only a few islands of small areas or few small areas in general are sampled. This may explain the great variety of relationships reported for and models fitted to the SIE (Lomolino and Weiser 2001, Triantis et al 2003, Dengler 2010, Triantis and Sfenthourakis 2012, Chisholm et al 2016, Schrader et al 2019. Our results therefore highlight that great care needs to be taken when collecting and analysing data for studying the SAR in general and the SIE in particular.…”

“…as many subtransects as possible on islands at their longest extension ( Fig. 1b, for more information on the study design see Schrader et al 2019). We inventoried all woody plants with a diameter at breast height ≥2 cm rooted within the transects.…”

“…Lomolino (2000) suggested that sigmoidal models are better suited as they test for the presence of the SIE while allowing a smooth transition between the range of the SIE and the linear phase of the SAR. Smooth transitions often provide more realistic representations of natural phenomena than sharp transitions (Toms and Lesperance 2013), such as the SIE (see also Chisholm et al 2016, Schrader et al 2019. Most authors agree that multiple alternative SAR models should be fitted and compared (Lomolino andWeiser 2001, Dengler 2010).…”

Species–area relationships on small islands differ among plant growth forms

“…Hence, the SIE prevails on islands where species richness is not governed by a dynamic equilibrium between colonization and extinction events, and the SIE should be more pronounced on more isolated islands that have lower colonization rates than on less isolated islands. Alternative, but not mutually exclusive, explanations for understanding the SIE stress the importance of limited habitat availability (Sfenthourakis & Triantis, 2009;Triantis, Mylonas, Lika, & Vardinoyannis, 2003), the area requirements of species (Schrader, König, Moeljono, Pärtel, & Kreft, 2019), greater influence of disturbances on smaller islands (Whittaker, 1995), alternating effects of niche-based and neutral processes (Chisholm, Fung, Chimalakonda, & O'Dwyer, 2016), or differential effects of isolation, topography and geology (Ackerman, Trejo-Torres, & Crespo-Chuy, 2007;Triantis et al, 2006).…”