Contrasting spatial pattern and pattern-forming processes in natural vs. restored shrublands

Abstract

Variation in the spatial arrangement of plants can lead to differences in the rates and trajectories of change in the composition, structure and function of plant populations and communities. While the ecological significance of spatial interactions has proven difficult to unravel in natural communities, restored ecosystems provide a simplified context for studying this problem. Examining spatial pattern in restoration is therefore useful for both restoration management and ecological theory. We compared 87 plant species patterns at two sites restored after sand-mining in southwestern Australia with 233 patterns at four nearby, species-rich, natural shrubland communities, examining a total of 193 species. Spatial tests were performed for all stems, and for all species with ≥20 individuals in each site, using Ripley's K and pair correlation functions with three null spatial models to characterize observed patterns. The processes and time-scales creating spatial pattern differ between restored and natural vegetation, so we hypothesized that, relative to natural sites, restored sites would have fewer gaps and less aggregation among all stems, fewer species with aggregated patterns, and a lower ratio of cluster- to gradient-aggregated patterns. We also hypothesized differences in pattern between species groups defined by seed size, dispersal, seed bank and regeneration traits, and that these would vary between natural and restored sites. However, relative to natural sites, restored vegetation had lower stem densities, disproportionately higher gap cover and similar patch attributes and mean nearest-neighbour distances. Most species were aggregated, and the fraction of aggregated species in restored sites fell within the range of natural sites. The frequency of aggregation varied little between species groups in restored sites, but strongly in natural sites, being rarer among serotinous and resprouting species.5. Synthesis and applications. Our spatial results show an important role for gap-creating processes in shrubland restoration development; that particular species groups may be susceptible to disproportionate decline under restoration; and how pattern-creating processes in natural vegetation vary with plant traits. Recommendations for restoration practice include compensatory supplementation of these species groups, research into the development of gaps and awareness of the potential for spatial pattern to influence outcomes.