Comparison of Forest Stand Edges in Riparian and Mesic Habitats Along Watts Bar Reservoir Shoreline

Approximately fifty years ago, the landscape upslope from the natural, riverine position banks of the Tennessee River was inundated by the closing of Watts Bar Dam. On the currently forested shoreline, habitats within the direct influence of the reservoir pool are riparian. Those above that influence are mesic. The purpose of this research was to determine compositional and structural differences between edges of mature forest stands established in a riparian habitat and those established in a mesic habitat along Watts Bar Reservoir shoreline. Thirty quadrats were placed on shoreline sites with mature, minimally-disturbed forest stands: 15 in a riparian habitat and 15 in a mesic habitat. Riparian and mesic habitats were distinguished by the hydric influence of the depth-to-subsurface lateral pool flow. A habitat was identified as riparian if subsurface lateral pool flow was estimated to be less than 0.5 m to soil surface (i.e., a low-lying area) and mesic if greater than or equal to 0.5 m (a topographically-elevated area). Each quadrat was 4 m wide x 25 m long and was located along the pool with the lengthwise edge being the summer pool line. Forest stand characteristics that were compared included vegetation structure (e.g., basal area, canopy height, canopy and edge closures) and composition (e.g., species diversity and richness, species importance values). Nonparametric statistics were employed for this comparison with supporting data provided by Two Way Indicator Species Analysis (TWINSPAN}, a clustering technique. Detrended Correspondence Analysis (DCA), an ordination technique, was further employed to determine whether any predominant underlying environmental gradient could be detected among the quadrats based on canopy species distribution. Results showed that sampled stands in riparian and mesic habitats were similar in productivity based on basal area, but differed significantly in their structure and composition. Stands in the mesic shoreline habitat exhibited characteristics of unmanaged broadleaf mesic forests. Twenty-nine hardwood taxa were represented in the canopy with a predominance of oaks and hickories. Stands in the riparian shoreline habitat were compositionally similar to regional bottomland forests and were limited to 16 canopy species with a predominance of Acer saccharinum. An assessment of similarity in canopy species of the two habitats yielded a Coefficient of Community of 0.33. The arboreal community in the mesic habitat was also significantly richer and more diverse than the community in the riparian habitat. TWINSPAN and DCA confirmed compositional dissimarilarity in sampled habitat stands. TWINSPAN partitioned mesic and riparian quadrats into two separate clusters. DCA segregated quadrats by habitat along an underlying environmental gradient. Analysis of this gradient indicated that it was related to subsurface lateral pool flow. Separate DCA analyses of mesic and riparian quadrats showed no predominant environmental gradient within either habitat. Structurally, riparian stands were significantly shorter, more open in their canopy, and denser in understory and edge front than mesic stands. Riparian stands characteristically presented a dense curtain-like edge cover composed of three common understory species, Cornus amomum, Alnus serrulata, and Ligustrum sinens