On the interaction between tree canopy position and environmental effects on soil attributes and plant communities

Questions: Understanding how trees affect their understorey plants and soils is crucial to understand savanna ecosystems. Most studies focus on the differences between canopy and open microsites, but how do different positions within large tree canopies influence soils and plants? Are these potential differences likely to change depending on environmental conditions (i.e. annual rainfall and grazing)? Location: One hundred sites across a rainfall gradient (220-1400 mm) in NSW, Australia. Methods: We measured the cover, richness and community composition of understorey plants and 12 soil attributes related to infiltration, erodibility and fertility across three positions within the canopy of large eucalypts (trunk, mid-canopy and edge) and in open areas. We also estimated the percentage similarity in plant communities across microsites, and the percentage species within the landscape occurring solely in one of the four microsites assayed. We tested the interactions between the effect of environmental conditions (rainfall and grazing) and canopy position on all these soil and vegetation attributes. Results: Soil attributes explained ~50% of the effect of trees on understorey plants, and soil attributes improved with proximity to the trunk and increasing rainfall. The effect of canopy position × rainfall interactions depended on the response variable considered. These interactions did not affect soil attributes, the percentage of facilitation-obligate species or species richness, and weakly affected plant composition. However, we found a strong reduction in similarity among plant communities within edge and mid-canopy compared with open sites towards drier environments, and canopy position × rainfall interactions also significantly affected plant cover. We attribute these weak interactions between canopy position and environmental conditions to richness or the frequency of facilitation to the high turnover of facilitated species across microsites and across different environmental conditions. Conclusions: Our study can be used to better understand community dynamics in ecosystems with scattered trees by showing the differential effects of trees on their understorey. Our results also contribute to the body of research on the relationships between plant-plant interactions and the environment by illustrating the importance of gradient length and the number of different microsites considered.