The role of population source and dominance on plant community diversity and productivity in a prairie restoration

Background/Question/Methods

Tallgrass prairie restoration provides an opportunity to study the effects of biotic filters on community and ecosystem reassembly.  Filters on prairie restoration include the source of dominant grass seeds (i.e., cultivars or local ecotypes), and the initial seed mix composition.  Our objective was to determine whether dominant grass source, and seeded grass dominance affect community composition and aboveground net primary productivity (ANPP) during prairie restoration.  We hypothesized that communities restored with cultivars and higher initial seeded dominance of grasses (i.e., higher grass:forb) would have lower diversity than communities restored with local genotypes and a greater proportion of forbs.  To test this hypothesis we established two restoration experiments in the western (Kansas) and eastern (Illinois) tallgrass prairie regions.  At both sites, we used a split plot design with source of the dominant grasses as the whole plot treatment and seeded diversity as the subplot treatment. 

Results/Conclusions

In Kansas, ANPP of the grass cultivars and non-cultivars was 77 and 18 g/m²/yr, whereas ANPP of grass cultivars and non-cultivars in Illinois were 253 and 274 g/m²/yr, respectively.  Planted forb ANPP in cultivar and non-cultivar grass treatments was 2 and 7 g/m²/yr in Kansas, and 56 and 67 g/m²/yr in Illinois.  Cultivars exhibited higher ANPP only in Kansas (P=0.019).  Subplot seeded dominance treatments affected the ANPP of dominant grasses (P=0.004) and planted forbs (P<0.001) exclusively in Illinois.  In Kansas, dominant grass cover in cultivar and non-cultivar plots was 18% and 10%, and planted forb cover was 7% and 5%, , respectively.  In Illinois, dominant grass cover in cultivar and non-cultivar plots was 40% and 33%, and planted forb cover was 31% and 35%, respectively.  Diversity of planted species was greatest in the low grass:forb seeding treatment at both sites (P<0.001).  In Kansas, diversity of planted species in cultivar plots (H'=1.41) was greater than diversity in non-cultivar plots (H'=1.12), (P<0.001).  In Illinois, planted species diversity was similar in cultivar plots (H'=1.39) and non-cultivar plots (H'=1.30), (P=0.310).  After two years of reassembly, dominant grass source and seed composition both affected structure and function in restored prairie, but not consistently across sites.  Population source effects were limited to Kansas, where conditions were drier; suggesting the potential assembly filter of seed source may be affected by abiotic conditions.  Seeded diversity responded similarly to dominance treatments across sites, suggesting propagule diversity was an important community assembly filter across the precipitation gradient.