Wednesday, August 4, 2010

PS 49-6: Effects of common disturbances on early-successional plant communities on Eleuthera Island, Bahamas

Claire C. Larkin1, Charles Kwit1, Joseph M. Wunderle Jr.2, Eileen H. Helmer2, M. Henry H. Stevens1, Montara T.K. Roberts3, and David N. Ewert4. (1) Miami University, (2) USDA Forest Service, (3) University of Minnesota, (4) The Nature Conservancy

Background/Question/Methods

Fundamentally different disturbances in similar habitats should produce unique successional assemblages of plants. We collected plant species composition data to investigate the effects of three common disturbances–fire, clearing (i.e., bulldozing), and clearing followed by goat grazing (‘cleared-and-grazed’)–on the vegetation composition and development of early-successional coppice (dry evergreen forest) patches (> 1 ha) on Eleuthera Island, Bahamas.  We expected (1) general differences in coppice plant communities as a function of disturbance type and (2) temporal differences in (a) subsets of species or species groups (e.g., species with low palatability such as Zanthoxylum fagara and Lantana spp.) and (b) total cover (%) of vegetation as a function of disturbance type. Using the line intercept method, both the ground layer (< 0.5 m height) and the shrub/tree layer (> 0.5 m height) were sampled in 8 patches of each disturbance type in which patch age (1-28 years) varied across a successional chronosequence.

Results/Conclusions NMDS ordinations based on percent cover showed separation in shrub layer communities between fire patches vs. cleared and cleared-and-grazed patches. Of the 79 species that were identified in the shrub layer for this study, percent covers of only three of these species including Acacia choriophylla, Bursera simaruba, and Coccoloba diversifolia were significantly different (P < 0.01) among the three disturbance types, all with highest cover in fire patches. Analysis of covariance showed B. simaruba and C. diversifolia percent cover increased along the chronosequence at different rates among the disturbance types (P < 0.01); quickest rates occurred in fire plots, followed by cleared and, lastly, cleared-and-grazed. Lantana spp. and Z.  fagara exhibited an initial increase in shrub layer cover, followed by a decline through the chronosequence. This general non-linear pattern, which conformed to a Ricker model, differed among patches with different disturbance types for Lantana spp. and Z. fagara (P < 0.01).  Total percent cover in the shrub layer linearly increased across the investigated chronosequence, and the rate of increase was similar across all disturbance types.  Our results suggest that goats may slow the increase in cover of major tree species. Goats may also allow non-palatable species such as Lantana spp. to thrive and remain within disturbed patches longer. This may have conservation implications for the Kirtland’s Warbler (Dendroica kirtlandii), which relies on fruits of Lantana spp. as the main source of food on its wintering grounds.