Seed bank dynamics under Rhododendron maximum: Implications for restoration of southern Appalachian forests

Background/Question/Methods

Riparian forest ecosystems once dominated by eastern hemlock (Tsuga canadensis) have undergone fundamental changes in function and composition from infestations by hemlock woolly adelgid. Eastern hemlock and Rhododendron maximum often co-occur in these riparian forests, with rhododendron forming a dense continuous subcanopy that inhibits seedling recruitment and alters forest community structure. Consequently, removal of rhododendron may be critical to restoring riparian areas affected by eastern hemlock loss. To investigate the potential contribution of the soil seed bank as a recruitment source in these degraded areas, we characterized the size and composition of the seed bank in riparian forests with and without a rhododendron subcanopy. In addition, we examined how changes in microclimatic and edaphic conditions beneath rhododendron may affect seed bank dynamics and the potential for seed germination. In June-July 2014, we measured vegetation, soil water content and temperature, photosynthetically active radiation (PAR), and forest floor depth in hardwood plots with and without rhododendron using a paired-plot design. Soil samples were collected in July 2014 and the seedling emergence method was used to estimate the density and composition of viable seeds in the seed bank. 

Results/Conclusions

We found few species in the soil seed bank within rhododendron thickets relative to open forests. Rubus was the most abundant species in the seed bank under rhododendron, while herbaceous species were exceedingly rare. Other woody species found in both forest types were primarily Liriodendron and Acer. Soil water content and temperature were significantly lower (P < 0.02) in plots with rhododendron than plots without rhododendron. PAR was relatively low (118 ± 7.9 µmol m^-2 s^-1) in these closed canopy forests with no significant difference (F = 0.57, P = 0.45) between the two forest types. Forest floor depth was greater (F = 64.5, P < 0.0001) in plots with rhododendron (75 mm) than those without rhododendron (21 mm). The buildup of the forest floor under rhododendron is due to the accumulation of recalcitrant leaves, which decompose slower than hardwood leaves. Our results suggest that the soil seed bank may not be the primary mode of recruitment to establish a diverse vegetative community when rhododendron is removed from these forests.