Assoc. Prof. of Biology Oberlin College Oberlin, Ohio, United States
Seed banks represent an often unrepresented component of plant diversity. General seed bank surveys are typically limited in their contributions to ecological analyses. We surveyed seed bank diversity with respect to habitat and disturbance patterns. Our study area is an 8 ha forest preserve in NE Ohio that has experienced the loss of all canopy ash trees (Fraxinus spp.) and invasion by multiflora rose (Rosa multiflora). The preserve has upland, slope, and floodplain habitats. We sampled three conditions (RoseOnly - high multiflora rose; RoseAsh - high ash loss w/multiflora rose; and Control - natural controls that lacked both ash and multiflora rose) in upland and floodplain habitats. We separated sample cores into the upper and lower 9 cm, with six replicates in each category for 84 samples. Following 5-week cold (4°C) treatments, soil samples were treated with gibberellic acid and spread over sterile soil, then grown under 14h: 25° days / 10 h: 15° nights for 30 days. We compared differences in germination and taxa patterns using Kruskal-Wallis, Mann-Whitney, and Χ2 tests, and with ecological diversity indices (species richness: Hill’s q=0, and effective number of species: Hill’s q=1). We predicted that disturbances increase diversity based on a secondary succession hypothesis.
Total germination produced 581 seedlings (mean=7.0 ± 9.4 s.d.; range=0-52). There were significantly (p< 0.05) more seedlings in the upper 9 cm, with ~2.4x as many as in the lower 9 cm. Highest abundances were Upland RoseAsh with 98 seedlings and floodplain RoseAsh with 88 (n.s.). Slope samples were lowest, with only 21 seedlings and no difference between upper and lower soil samples. We identified 88 taxa, eleven to the species level, with marginally significant differences among the 14 treatments (p=0.055). We pooled the six samples for each category, which showed that taxa richness varied from 6 (slope, lower 9cm) to 19 (RoseAsh Floodplain upper 9 cm) with mean=10.3 ± 3.9 s.d. taxa per treatment. Effective number of taxa varied from 4.14 (slope, lower layer) to 11.03 (RoseAsh Floodplain upper layer) with mean=5.8±1.7 s.d. We found a significant difference (p< 0.05) between richness and effective number when comparing the upper (mean=6.0±2.4 s.d.) and the lower 9cm (mean=3.1±1.0 s.d.) samples. The upper 9cm of RoseAsh sites had the highest abundance (p< 0.05) and richness (n.s.) in both upland and floodplain sites. RoseOnly and Control sites were approximately equal. Our results show that ash loss is the primary disturbance factor.