SEARCH

SEARCH BY CITATION

Keywords:

  • Coastal sand dunes;
  • Competition;
  • Disentangling environmental stresses;
  • Facilitation;
  • Fertilization;
  • Stress gradient hypothesis;
  • Water availability

Abstract

Questions

What is the relative role of water vs nutrients in driving changes in plant interactions in a dry and nutrient-poor dune system, where facilitation has been shown to highly contribute to the diversity and composition of plant communities?

Location

Atlantic coastal sand dune system, Aquitaine region, France.

Methods

We manipulated water (water addition) and nutrient (fertilizer addition) availabilities in a full-factorial experiment that lasted from September 2005 to September 2006. The responses of plants to neighbours were assessed through transplanting six target species within each experimental plot in the presence or absence (removal procedure) of the dominant shrub Helichrysum stoechas. We also measured changes in H. stoechas biomass among treatments.

Results

Watering mainly affected survival of the target species without neighbours and the direction of their responses to the shrub, whereas fertilization strongly increased the biomass of the shrub and intensity of the responses of the targets to the effects of the shrub. We did not observe any occurrence of biotic interactions (either positive or negative) for most of the target species in the unfertilized plots. In contrast, three of the six species exhibited significant facilitation in fertilized and un-watered plots, and significant competition in fertilized and watered plots.

Conclusions

Our results highlight the co-limitation of water and nutrients in coastal sand dunes. Nutrient availability mainly drove the neighbour effect of the shrub, whereas water availability affected the responses of the target species. Although further measurements of resource modifications by the shrub are needed to firmly conclude neighbour effects, our results provide new insights to the contrasting results in the literature on plant–plant interactions in unproductive communities. We propose that, depending on the system, changes in a given direct stress factor may either induce a collapse of interactions by decreasing neighbour performances and their effects, or a shift in the direction of interactions by affecting target species responses.