Evaluating the success of bottomland forest restoration in the Upper Mississippi Valley

Background/Question/Methods

Bottomland forest (BLF) provides vital ecosystem services (e.g., water quality, nutrient cycling, wildlife habitat, flood, and sediment control). Since European settlement, clearing for timber and agriculture and changes in hydrology have significantly reduced the area of BLF in the Upper Mississippi Valley (UMV). Beginning in 1990s, the US Army Corps of Engineers has undertaken BLF restoration on many sites in the UMV. Root production method seedlings of mainly large-seeded species (bottomland oaks and pecan), which are not regenerating under current conditions, have been planted with the expectation that light-seeded species (e.g., silver maple, green ash) will colonize passively. A chronosequence of nine restoration sites, ranging in age from 1 to 23 yr, and two mature BLF reference sites was used to assess restoration success. Five 0.1 ha circular plots were randomly located at each site. Planted trees and natural recruits with DBH ≥ 2.5 cm were identified, tagged and their basal diameter, DBH, and height measured. Density of shrubs was assessed in belt transects and cover of herbaceous species was estimated in 0.5 m²quadrats. Trajectories of change in tree size, community structure, and diversity were examined using generalized linear modeling, relative to their values in reference sites.

Results/Conclusions

Tree height and diameter increased with time since restoration for all species. Quercus palustris, Q. macrocarpa, and Q. bicolor are all on track to achieve dimensions typical of mature BLF forest. Pecan has suffered high mortality and without replanting it will be underrepresented in the restored forest. Tree richness increased slowly with age as light-seeded species have supplemented the planted species but the oldest restoration sites are still lower in richness than reference sites. Both total shrub density and total herbaceous cover showed a slow decline over time, although diversity of both shrubs and herbaceous vegetation increased. Comparison with reference sites suggest that the shrub and herbaceous layers will continue to decline in abundance with development of a forest canopy and the diversity of these layers will peak and then decline. Overall, our results indicate restoration success but suggest that replanting will be necessary in most sites to overcome mortality due to prolonged flooding and other factors (e.g. white-tailed deer browsing).