Land-use history, including disturbances and non-native species introductions, can control the long-term trajectories of plant communities and ecosystem processes. We hypothesized that persistent effects on soil microbial communities are a mechanism underlying such land-use legacies. Microbial legacies include loss of keystone species or other compositional changes that can affect microbe-driven ecosystem processes. We addressed the persistence of soil microbial legacies and their remediation using a large-scale restoration experiment in a xeric Florida shrubland. In the native scrub ecosystems, biological soil crusts dominate the soil surface and play critical roles in nitrogen cycling, soil water retention, and seed germination. The restoration experiment was focused on areas that had been subjected to different degrees of non-native grass introductions and grazing management, resulting in ‘disturbed scrub’ and ‘converted pasture’ sites. We removed non-native grasses and reintroduced native soil crusts in a factorial design, with both untreated controls and native undisturbed reference sites. Because of their likely keystone role in nitrogen fixation, we tracked the recovery of cyanobacteria and examined how that recovery related to nitrogen cycling over two years. We examined cyanobacteria using chlorophyll fluorescence and 16S rDNA. Soil inorganic nitrogen was extracted with KCl and nitrogen fixation was estimated with acetylene reduction.

Results/Conclusions

Cyanobacteria and soil nitrogen were similarly affected by the degree of disturbance, with the greatest abundances of cyanobacteria and the highest concentrations of nitrogen in converted pasture soils followed by disturbed scrub and native sites. The effectiveness of the restoration treatments was also site-specific. In converted pastures, only the combination of non-native plant removal and crust inoculation successfully reduced cyanobacteria to levels found in native sites over two years. In disturbed scrub sites, however, non-native plant removals were sufficient for a transient impact (1-year) while the addition of crust inoculation extended this effect for a second year. Treatment effects on soil nitrogen paralleled those of cyanobacteria abundance, suggesting that the cyanobacteria community supported in pasture sites may contribute to the persistence of an altered nitrogen cycle. Preliminary DNA analysis suggests that differences in cyanobacteria community composition are consistent with this role. Our results support the hypothesis that soil microbial communities act as drivers of ecosystem legacies. Directly targeting soil microbial communities for management broadens our restoration toolkit and may help increase the success of restoration efforts.