NIU Graduate Student Research: Systematic Evaluation of Prairie Planting Regimes

Conservation ecologists over the past several decades have sought to better understand the factors that promote biodiversity in and the resilience of ecological systems. In order to forward this effort, Brian Glaves, an NIU Geography graduate student, engaged in an experiment to identify the optimal amount or density of native seed required to produce a successful prairie planting.

Rough Blazing Star (Liatris aspera) blooming within Nachusa Grasslands

Brian's project was situated within the idyllic 3,000 acre Nachusa Grasslands, a conservation area in north central Illinois. The project was an extension of a previous experiment carried out in 2006 by Bill Kleiman who, along with numerous volunteers, manages the grasslands. For this earlier experiment, Kleiman applied a dry mesic mix of native Nachusa Grasslands seeds to a smaller region of the conservation area known as Clear Creek Knolls. This seed density experiment established a 3 x 5 systematic random quadrant design in which one of four categories of seed density treatments (10, 30, 50 and 70 pounds per acre) were applied or dispersed. The seeding was replicated three times with each treatment assigned randomly to a (9m x 9m) quadrant in order to control for local variation. The systematic random quadrant design provided a standard method for examining and recording the performance of each treatment relative to each other and a control group.

In the summer months of June, July and August 2008, Brian began a systematic recording of floristic data from the 2006 planting. Specifically, Brian measured species richness, species density, percent flora cover and species height. A variety of indices such as the Shannon-Weiner diversity index (i.e., frequency and evenness of each species within the community), Simpson's dominance index (i.e., a measure of the dominance of any one species) and a Floristic Quality Index were then used to characterize the data collected in the field and examine the relative performance of each seed density regime.

Brian's results showed no statistically significant difference between the various indices across the four categories of seed density treatments, although some general trends were apparent. For example, lower seed density treatments tended to exhibit higher Simpson's (dominance) values (Figure on left), while greater densities exhibited higher Shannon-Weiner (diversity) values (Figure on right). The Forensic Quality Indices also suggest that treatments were significantly different from the control (One-way ANOVA, p=0.05) but not to each other. Further, the total plant cover varied over the study periods such that plant cover as observed in June was statistically different from the plant cover observed in July and August. This was in part due to the greater abundance of exotic species earlier in the season. Lastly, the data suggest that individual species tended to be better represented after applying the 50 pounds per acre seed density treatment. In fact, the statistical analyses suggest that the 50 pounds per acre seed density treatment (rather than the more resource intensive 70 pound treatment) was the best option for Nachusa Grasslands prairie restoration while the 30 pounds per acre treatment was acceptable.

Summer Crew at Nachusa Grasslands, 2008
Brian Glaves (front, right), Bill Kleiman-Manager of Nachusa Grasslands (left of Brian Glaves)

So why is this research important? Most research to date on seed density applications for prairie restoration has been rather informal in its analysis of planting regimes. Brian's research puts forward a more scientifically rigorous strategy for evaluating successful prairie planting. In particular, what this experiment showed is that prairie restoration seeding practices can be defined within rather clear ranges, in this case 30 to 50 pounds of seed per acre.