The seed bank of a restored tidal freshwater marsh in Washington,DC

The species composition of seed banks may be a useful indicator of vegetation structure and dynamics, but seed bank studies have rarely been applied to monitoring of wetland restoration projects. We studied the seed banks and vegetation of three distinct sites within a restored tidal freshwater marsh and of two nearby natural marsh sites. Surface soil samples were collected and subjected to flooded and nonflooded conditions in a greenhouse for enumeration of emerging seedlings. Percent coverage of vegetation was estimated in permanent 1-m² plots at each soil sampling location.Species composition of the seed bank and vegetation differed considerably; only 43% of the species observed were common to both. The seed bank of restored marsh sites had higher species richness and seed density than the natural marsh sites, but both had similar species diversity. Annual species were more important in the seed bank and less important in the vegetation of restored sites than reference sites. Flooding significantly reduced the species richness, diversity, and density of emerging seedlings. Most species planted at the site were rare in the seed bank, with Leersia oryzoides being an important exception. Other species abundant in the seed bank included Cyperus odoratus, Juncus effusus, and the nonindigenous Lythrum salicaria. A possible mechanism explaining the higher densities of seeds in restored areas is that ruderal vegetation colonized the restoration site, reproduced, and set seed shortly following placement of dredge material. Perennial species subsequently colonized and were planted, and expanded to reduce the abundance of annuals in restored marsh vegetation. Our study suggests that the plant regeneration dynamics of the restored marsh sites are considerably different from those of natural marshes. Additionally, we suggest that seed bank studies are a useful technique for assessing the community structure and vegetation dynamics of restored wetlands.