Restoration barriers of novel grassland ecosystems in southeastern Australia

Background/Question/Methods

Restoring ecological integrity to native grasslands following years of intensive agricultural use is integral to achieving scientific and social goals from maintaining biodiversity to creating natural spaces for recreation. A large grassland reserve has recently been created from retired pasture on the outskirts of Melbourne, Australia. This grassland is dominated by the native grazing-tolerant grass Austrostipa bigeniculata, with low native forbs diversity, while remnant patches of biodiverse Themeda trianda-dominated grassland provide an indication of the historical state and a clear restoration target.

With over 13,000 hectares to be restored, it is essential to determine the most efficient pathway to restoration. The objective of this study was to identify the biotic and abiotic controls on community assembly with a focus on native forb establishment, and to predict vegetation trajectories given different restoration interventions. First we assessed the soil chemistry across the different vegetation states in the reserve. We then explored biotic controls on native forb establishment by conducting a seeding experiment following a natural fire. Lastly, we used Bayesian models to predict the impacts of different interventions on community composition. 

Results/Conclusions

The Themeda grasslands had entirely different soil chemistry from the retired pasture sites, with Themeda sites differentiated largely by their lower phosphorus concentration, higher pH, and lower aluminum concentration. Eight of the ten planted native forb species successfully established in newly burnt Austrostipa grasslands although it was not clear that all species would form persistent populations. This establishment occurred regardless of whether soils were additionally disturbed by raking when seeds were sown, although disturbance did improve establishment rates for some species. Total grass cover decreased with an increase in planted native forb cover. However, this appeared to be driven by the soil disturbance in the raked treatments rather than by seed competition.

Altogether these results suggest that while some species may be absent because of low propagule pressure, restoration barriers may be substantial for many of the native forb species. These results highlight the need for flexible restoration goals for highly modified novel ecosystems.