Aim: To disentangle the influence of environmental factors at different spatial grains (regional and local) on fern and lycophyte species richness and to ask how regional and plot-level richness are related to each other. Location: Global. Taxon: Ferns and lycophytes. Methods: We explored fern and lycophyte species richness at two spatial grains, regional (hexagonal grid cells of 7,666 km2) and plot level (300–500 m2), in relation to environmental data at regional and local grains (the 7,666 km2 hexagonal grid cells and 4 km2 square grid cells, respectively). For the regional grain, we obtained species richness data for 1,243 spatial units and used them together with climatic and topographical predictors to model global fern richness. For the plot-level grain, we collated a global dataset of nearly 83,000 vegetation plots with a surface area in the range 300–500 m2 in which all fern and lycophyte species had been counted. We used structural equation modelling to identify which regional and local factors have the biggest effect on plot-level fern and lycophyte species richness worldwide. We investigate how plot-level richness is related to modelled regional richness at the plot's location. Results: Plot-level fern and lycophyte species richness were best explained by models allowing a link between regional environment and plot-level richness. A link between regional richness and plot-level richness was essential, as models without it were rejected, while models without the regional environment-plot-level richness link were still valid but had a worse goodness-of-fit value. Plot-level richness showed a hump-shaped relationship with regional richness. Main conclusions: Regional environment and regional fern and lycophyte species richness each are important determinants of plot-level richness, and the inclusion of one does not substitute the inclusion of the other. Plot-level richness increases with regional richness until a saturation point is reached, after which plot-level richness decreases despite increasing regional richness, possibly reflecting species interactions.
Bibliographical noteFunding Information:
We are grateful to the numerous people who have contributed to the data used in this paper by participating in field work, helping with practical arrangements for the field expeditions, data management or sharing their expertise for species identification. We thank the national authorities in each country for granting the permits to carry out field work and collect voucher specimens. Funding and help that have made this work possible have been provided by numerous agencies and foundations over the years, including the Academy of Finland, Finnish Cultural Foundation, FAPESP/FAPEAM/FAPESC, CNPq, the Brazilian Program in Biodiversity Research (PPBio), Canadian Forest Service (Natural Resources Canada), Ontario Ministry of Natural Resources and Forestry, SERG International members, Natural Sciences and Engineering Research Council of Canada, the Institute of Biology Bucharest of Romanian Academy (RO1567‐IBB01/2019), and The Rufford Foundation (Honduras). The publication also contributes as publication 775 to the technical series of the Biological Dynamics of Forest Fragments Project. The sPlot project was initiated by sDiv, the Synthesis Centre of the German Centre for Integrative Biodiversity Research (iDiv) Halle‐Jena‐Leipzig, funded by the German Research Foundation (DFG FZT 118) and is now a platform of iDiv. We gratefully acknowledge the Programme Coordinating Centre of ICP Forests and all botanists who performed the vegetation surveys. The evaluation included data from the UNECE ICP Forests PCC Collaborative Database (see http://icp-forests.net ). Data from the following participating countries (number of plots in parentheses): Austria (177), Belgium (351), Bulgaria (18), Cyprus (16), Czech Republic (47), Denmark (69), Estonia (110), Finland (121), France (6768), Germany (1329), the United Kingdom (390), Greece (15), Hungary (279), Ireland (12), Italy (766), Latvia (11), Lithuania (18), Luxemburg (112), Netherlands (47), Norway (21), Poland (160), Portugal (58), Romania (37), Russia (147), Serbia (238), Slovenia (64), Slovakia (35), Spain (83), Sweden (166) and Switzerland (755) were included in this study. The collection of ICP Forests data was to a large extent funded by national research institutions and ministries, with support from governmental bodies, services and landowners and the European Commission. We further thank the anonymous reviewers and the editors for helpful comments.
- big data
- regional-local richness relationship
- saturation curves
- structural equation modelling