Conditional love? Co-occurrence patterns of drought-sensitive species in European grasslands are consistent with the stress-gradient hypothesis.

Stephan Hennekens,Luca Santini,Aafke M. Schipper,Melinda M. J. de Jonge,Ana Benítez‐López,Mark A. J. Huijbregts

doi:10.1111/geb.13323

Abstract

AimThe stress‐gradient hypothesis (SGH) postulates that species interactions shift from negative to positive with increasing abiotic stress. Interactions between species are increasingly being recognized as important drivers of species distributions, but it is still unclear whether stress‐induced changes in interactions affect continental‐to‐global scale species distributions. Here, we tested whether associations of vascular plant species in dry grasslands in Europe follow the SGH along a climatic water deficit (CWD) gradient across the continent.LocationDry grasslands in Europe.Time periodPresent.Major taxa studiedVascular plants.MethodsWe built a context‐dependent joint species distribution model (JSDM) to estimate the residual associations (i.e., associations that are not explained by the abiotic environment) of 161 plant species as a function of the CWD based on community data from 8,660 vegetation plots. We evaluated changes in residual associations between species for pairs and on the community level, and we compared responses for groups of species with different drought tolerances.ResultsWe found contrasting shifts in associations for drought‐sensitive and drought‐tolerant species. For drought‐sensitive species, 21% of the pairwise associations became more positive with increasing CWD, whereas 17% became more negative. In contrast, only 17% of the pairwise associations involving drought‐tolerant species became more positive, whereas 27% became more negative in areas with a high CWD. Additionally, the incidence of positive associations increased with drought for drought‐sensitive species and decreased for drought‐tolerant species.Main conclusionsWe found that associations of drought‐sensitive plant species became more positive with drought, in line with the SGH. In contrast, associations of drought‐tolerant species became more negative. Additionally, changes in associations of single species pairs were highly variable. Our results indicate that stress‐modulated species associations might influence the distribution of species over large geographical extents, thus leading to unexpected responses under climate change through shifts in species associations.

Highlights

It has long been assumed that species distributions at large spatial scales are determined mainly by abiotic factors (Wisz et al, 2013)
The community association (CA) for drought-sensitive species increased with the climatic water deficit during the growing season (CWDGS) in the second half of the gradient, that is, there were more positive compared with negative associations in dry than in wetter environments, which is in line with the expectations from the stress-gradient hypothesis (SGH) (Figure 4b)
Species distributions have been considered to be determined mostly by abiotic factors, whereas our results suggest that biotic interactions play a role and that these interactions might change along environmental gradients

Summary

| INTRODUCTION

It has long been assumed that species distributions at large spatial scales are determined mainly by abiotic factors (Wisz et al, 2013). Disentangling the relative importance of species interactions and abiotic factors is challenging, because abiotic conditions might alter the intensity and direction of the interactions (Bertness & Callaway, 1994; He et al, 2013) It is still unclear whether such environment-dependent shifts in species interactions influence the continental-to-global scale distribution of species (Early & Keith, 2019; Wisz et al, 2013). Species with different life-history strategies can show opposite responses to their neighbours, which might obscure community-level patterns (Graff & Aguiar, 2017; Michalet et al, 2015) It remains unclear whether the large-scale distributions of plant species are influenced by drought-modulated species interactions and whether these follow expectations of the SGH. We expected that drought-tolerant species would not be affected by the water deficit gradient, their associations would not change

| METHODS

Findings

| DISCUSSION