During landfall (2005), the Hurricane Katrina storm surge pushed seawater and floating debris several kilometers inland at the Grand Bay National Estuarine Research Reserve (Mississippi). Wrack bands of woody and herbaceous plant debris 1-2m thick buried groundcover vegetation in coastal savannas with slash pine (Pinus elliottii) in the overstory and a diverse groundcover dominated by cordgrass (Spartina patens). We conducted a field experiment to test hypotheses regarding effects of burial under and subsequent removal of wrack on survival and colonization by species of groundcover plants: 1) Species with different life forms differentially survive burial by wrack; 2) Removal of wrack differentially facilitates colonization by plants with different life forms. We randomly established 1 m² subplots, 8 in areas with and 8 in adjacent areas without wrack in each of three separated plots (total 48 subplots). We recorded species composition in subplots in June, 2008, and then manipulated wrack in half of the subplots in each plot. We simulated a new hurricane that shifted wrack deposits, exposing the soil substrate in subplots where wrack was removed, and depositing wrack in subplots without wrack. We resampled species composition in subplots one and two years (June 2009, 2010) after treatment.

Results/Conclusions

Adding wrack to subplots without wrack resulted in >50% reductions in species and shifts in dominant life forms from grasses and forbs to shrubs and lianas. Rosette life forms were most susceptible, while life forms that produce long slender culms/stems capable of extension through wrack deposits (e.g., Spartina patens, Ipomoea sagittata) or stout stems (e.g., Ilex vomitoria) were more resistant. We project that persistence should depend on amounts of wrack deposited; as hurricane intensity increases, fewer species should survive burial. Removing wrack from subplots resulted in variable, but three/five-fold overall increases in numbers of species, almost entirely graminoids and forbs not present in subplots without wrack. Low (5-10%) overlap in species composition occurred between wrack removed and no-wrack subplots after two years. Many groundcover graminoids and forbs in coastal pine savannas responded rapidly once wrack deposits no longer were present. Fugitive (or early) species with dormant seeds beneath wrack survived burial and rapidly colonized areas once wrack was removed, more than doubling local groundcover biodiversity. We project that periodic wrack disturbances of variable intensity should generate different local spatial combinations of early, resilient species and late, resistant species, resulting in high heterogeneity and thus overall diversity in coastal pine savannas.