Multitrophic diversity in a biodiverse forest is highly nonlinear across spatial scales

ARTICLE Received 24 Apr 2015 | Accepted 10 Nov 2015 | Published 10 Dec 2015 DOI: 10.1038/ncomms10169 OPEN Multitrophic diversity in a biodiverse forest is highly nonlinear across spatial scales Andreas Schuldt1, Tesfaye Wubet2,3, Franc¸ois Buscot2,3, Michael Staab4, Thorsten Assmann1, Martin Böhnke-Kammerlander5, Sabine Both6, Alexandra Erfmeier3,7, Alexandra-Maria Klein4, Keping Ma8, Katherina Pietsch9, Sabrina Schultze1, Christian Wirth3,9, Jiayong Zhang10, Pascale Zumstein1 & Helge Bruelheide3,5 Subtropical and tropical forests are biodiversity hotspots, and untangling the spatial scaling of their diversity is fundamental for understanding global species richness and conserving biodiversity essential to human well-being. However, scale-dependent diversity distributions among coexisting taxa remain poorly understood for heterogeneous environments in biodiverse regions. We show that diversity relations among 43 taxa—including plants, arthropods and microorganisms—in a mountainous subtropical forest are highly nonlinear across spatial scales. Taxon-specific differences in b-diversity cause under- or overestimation of overall diversity by up to 50% when using surrogate taxa such as plants. Similar relationships may apply to half of all (sub)tropical forests—including major biodiversity hotspots—where high environmental heterogeneity causes high biodiversity and species turnover. Our study highlights that our general understanding of biodiversity patterns has to be improved—and that much larger areas will be required than in better-studied lowland forests—to reliably estimate biodiversity distributions and devise conservation strategies for the world’s biodiverse regions. 1 Institute of Ecology, Leuphana University Lüneburg, Scharnhorst Strae 1, D-21335 Lüneburg, Germany. 2 Department of Soil Ecology, UFZ-Helmholtz Centre for Environmental Research, Theodor-Lieser-Strae 4, D-06120 Halle (Saale), Germany. 3 German Centre for Integrative Biodiversity Research (iDiv) Halle-JenaLeipzig, Deutscher Platz 5e, D-04103 Leipzig, Germany. 4 Institute of Earth and Environmental Sciences, University of Freiburg, Tennenbacher Strae 4, D-79106 Freiburg, Germany. 5 Institute of Biology/Geobotany and Botanical Garden, University of Halle, Am Kirchtor 1, D-06108 Halle, Germany. 6 Institute of Biological and Environmental Sciences, University of Aberdeen, St Machar Drive 23, AB24 3UU Aberdeen, UK. 7 Institute for Ecosystem Research, University of Kiel, Olshausenstrasse 75, 24118 Kiel, Germany. 8 Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China. 9 Systematic Botany and Functional Biodiversity, University of Leipzig, Johannisallee 21-23, D-04103 Leipzig, Germany. 10 Institute of Ecology, Zhejiang Normal University, Yinbing Road 688, Jinhua 321004, China. Correspondence and requests for materials should be addressed to A.S. (email: ). NATURE COMMUNICATIONS | 6:10169 | DOI: 10.1038/ncomms10169 | www.nature.com/naturecommunications 1 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10169 U ntangling the scale-dependency of a- and b-diversity among coexisting taxa is essential to understand the structuring of ecological systems, to estimate regional and global species richness, and to inform policy options on conservation1–6. However, how exactly megadiverse groups such as arthropods and microorganisms scale in relation to more frequently assessed taxa, such as plants, is a matter of ongoing debate. This particularly applies when extrapolating assessments to landscape and regional scales in the most speciesrich terrestrial regions of the world, subtropical and tropical forests2,7. Most studies on scale-dependent biodiversity patterns in species-rich forests have focused on single taxa8–13. Those studies that have considered multiple taxa have analysed various—including non-forest—habitat types or restricted the spatial analyses to pairwise plot comparisons14–17. Despite their functional importance, microorganisms have so far been ignored in such studies. A recent study in a lowland neotropical rainforest, however, showed similarities in species turnover (b-diversity) for a wide range of arthropod taxa7. By extrapolating local plot species inventories, that study showed that areas as small as 1 ha can harbour almost two-thirds of the landscape-scale species richness. Moreover, the species richness of arthropods across all trophic levels was surprisingly well predicted by that of woody plants, and this strong relationship was independent of the geographic scale considered7. However, whether these patterns can be extrapolated to species-rich forest types in more heterogeneous environments, and to other species-rich taxa such as microorganisms, is questionable. Many highly diverse forests, and many of the world’s biodiversity hotspots18, are located in mountainous landscapes with heterogeneous topography, which results in a higher b-diversity of many taxa than in more homogeneous lowland forests9,19. This may have consequences for the design and costs of biodiversity research and conservation, and for species richness estimates at larger spatial scales19,20. We conducted a comprehensive assessment of the species richness, turnover, and cross-taxon diversity congruence of plants, arthropods and, for the first time, soil microorganisms from the local plot to landscape scales in a highly diverse, and topographically and environmentally heterogeneous, subtropical forest. We used multi-method species censuses of above- and below-ground organisms (woody and herbaceous plants; 10 arthropod taxa comprising herbivores, detritivores, predators and parasitoids; 12 groups of soil fungi and 19 groups of bacteria) and modelled species richness and area relationships (Methods). The data were obtained from 27 study plots that reflect the environmental heterogeneity typically found in the 8,000-ha mountainous study site, a national forest reserve in South-East China. Our analysis shows that cross-taxon diversity relationships are highly nonlinear across spatial scales, with far-reaching consequences for our understanding of regional and global biodiversity patterns and their conservation. Results Sampling completeness. Altogether, we identified 1,008 (morpho)species of arthropods and plants with a total of 77,718 individuals, and 6,223 operational taxonomic units (OTUs) of microorganisms. Species–area relationships for all taxa were best modelled by asymptotic functions. Sample coverage, a measure of sample completeness, approached values 490% relatively fast with increasing plot number, and was on average 97% across all taxa in 27 plots (with the exception of lepidopterans; Supplementary Fig. 1). 2 Spatial scaling of species richness patterns. On average, 1 and 10 ha of the subtropical forest at our study site can be expected to capture 38% and 76%, respectively, of the overall estimated species richness for the 10 arthropod taxa, 71% and 97% for the 12 fungal taxa and 93% a (...truncated)