Native grass establishment through seed addition in a southeast Kansas oak savanna restoration

Background/Question/Methods

Globally, woody plant population expansion threatens savannas.  Restoring savannas from woodlands can involve tree thinning, but often the herbaceous understory is non-existent and potential seed sources may be distant.  In a blackjack oak – post oak savanna restoration in the Cross Timbers ecosystem of southeast Kansas, I examined whether 1) seed addition enhances native grass establishment and 2) spatial heterogeneity created by oak canopies modifies the relationship between seed addition and grass establishment.  In March 2009 after tree thinning earlier that winter, I sowed Indian grass (Sorghastrum nutans) and Switch grass (Panicum virgatum) seeds at four densities (0, 50, 250 and 1000 seeds) in 0.5 m X 0.5 m plots at oak canopy edges and between canopies.  

Results/Conclusions

In the third growing season after sowing (2011), the proportion of plots with S. nutans present was significantly greater under all levels of seed addition than without seed addition.  For P. virgatum, addition of 250 seeds and 1000 seeds / 0.5 m X 0.5 m plot resulted in a significantly higher proportion of plots with P. virgatum present than did no seed addition.  Nevertheless, cover and tiller densities of both grasses remained low (mean values <5.5% cover and <16 tillers / 0.5 m² for 1000 seeds added).   For Indian grass, sowing 1000 seeds increased cover and tiller density over no seed addition controls.  For Switch grass, sowing 1000 seeds increased cover, but not tiller density.  Lower sowing densities did not significantly increase cover or tiller density.  Plot position relative to oak canopies did not affect the grass species’ establishment.  While seed addition was effective in overcoming seed limitation of native grass establishment, any large differences in cover of the added native grasses between seed sowing densities are likely to develop gradually because clones of these grass species expand slowly.  High rates of seed addition likely produce greater cover through establishment of a larger number of genets.