Mechanisms of plant recovery following complete restoration in a small floodplain

Background/Question/Methods

Many streams and rivers in the US have had their basic functions altered by human manipulation.  Restoration of function in these systems requires intensive efforts.  A stream in central Wisconsin that was channeled in the mid-20^th century was restored in 2010 by complete removal of all woody vegetation, large-scale soil removal, creating meanders in the channel, and re-distribution of the topsoil.  In the fall, 2010, a wetland mix consisting of 14 graminoids and 14 forbs was seeded over the 6.9 ha project area. Late in the 2011 growing season, we established a systematic array of plots consisting of 103 quadrats along 24 transects.  The cover and identity of all plants was determined, and two small soil cores were taken adjacent to the quadrat.   The seed bank of the soil was determined by the germination method; half of each sample was stratified.

Results/Conclusions

At the end of the first growing season, the typical quadrat contained approximately  7 species (no exotics), had 37% vegetative cover and 55% bare mineral soil.  The assemblage was strongly dominated by sedges,  rushes  and grasses; these groups contributed 6 of the 8 most frequently occurring species.   An ordination indicated no clustering of plots, suggesting relatively consistent environmental conditions across the floodplain.  The origin of the species  was:  38% from the seed bank, 22% invaded from outside the project area, 6.5% were in the seed bank and in the sown mixture, but only 5.5% were solely from the species sown, and origin could not be definitely determined for 27.5%.  The latter group was about one-half  immature individuals identified to Carex , which  could have originated from the seed bank and/or the sewn mixture.  These results are striking because of the major contribution of the seed bank, despite having been moved and piled for > 1 month.  The seed bank contributed 5 species (out of 14 total) with strong wetland affiliation, whereas about 90% of the species that ‘invaded’ had similar association.   The contribution of all wetland species was 49% of the total cover.  The amount of cover from the seed bank plus ‘invader’ group  was negatively correlated (r = -.43) with bare mineral soil.  Exotic species made up a very small (< 1.0 %) of the cover and richness of the assemblage.  The community that has re-assembled has a strong wetland representation despite severe disturbance and a small contribution from the sewn species.