Accumulating evidence suggests that restoration projects, especially in hyper-diverse ecosystems, generate novel ecosystems rather than recreating a historical or reference condition. To guide management of novel ecosystems towards a close-to-reference goal, this study sought to understand how to apply environmental filter theory to reinstate a biodiverse local plant community on degraded post-agricultural land.
Three environmental filters: dispersal (topsoil seed bank density and seed dormancy), abiotic (topsoil compaction and moisture) and biotic (grazing and competition), were manipulated to examine their effect on the trajectory of community development towards the desired reference Banksia woodland ecosystem. Banksia woodland is a hyper-diverse ecosystem restricted to the Swan Coastal Plain in Western Australia that is diminishing due to rapid urban sprawl. Topsoil from under cleared Banksia vegetation that contains a large and diverse native soil seed bank, was salvaged and transferred to a 25ha restoration site. Then a fully factorial combination of filter manipulation treatments was applied across six sites to identify successful remediation techniques. The dispersal filter was tested by altering volume of topsoil and application of two fire-related treatments (smoke and heat). The abiotic filter manipulation was performed using topsoil ripping and plastic cover to enhance soil moisture content. The biotic filter was examined using herbivore exclosures and weed control. Emergence and survival of Banksia woodland species were quantified in spring and autumn for two consecutive years after topsoil transfer.
Results/Conclusions
Manipulation of the dispersal filter, by applying a larger volume of topsoil, proved to be the best in optimizing restoration of the native Banksia ecosystem. This treatment returned approximately 25% of native plant species found in the reference plant community of Banksia woodland (144 species records in this study). Application of deep topsoil (10cm) increased native seedlings emergence by 12% compared with application of shallow topsoil (5cm). Deep topsoil decreased weed densities by 25%. Overall seedling survival over the 2-year sampling period was low – 0.91% regardless of treatment-type, in part due to unusually dry weather conditions. The plant assemblages after two years comprised mostly of non-native perennial grasses and perennial, small-seeded native woody shrubs, but was missing the crucial canopy tree species as this community component is not represented in the soil seed bank. The final plant assemblages formed a novel type of ecosystem that, in part, meets completion criteria set by local policy makers but restoration of many of the desired community properties remains an elusive goal.