The competitive release hypothesis predicts that flood-tolerant plant species will occupy the wet end of a hydrologic gradient, but will be displaced by species with higher competitive ability at the upper end of the gradient. In this study, we tested this hypothesis in a meso-scale field experiment, examining species responses to flooding and neighbor size/proximity during Years 3-8 after initial planting. In 2002 a landscape-scale physical model of the Everglades, the Loxahatchee Impoundment Landscape Assessment (LILA), was constructed at Loxahatchee National Wildlife Refuge. Control of hydrology and replication of landform structure at LILA allowed investigators to precisely assess ecosystem responses to important physical or biological drivers. One component of the landscape created at LILA was a set of forest patches (tree islands) that show a moisture-elevation gradient from drier center to wetter edges. Mean water depth was estimated for each tree using elevation data from topographic surveys and water level data from nearby stage recorders. An interference index was used to characterize the competitive neighborhood of a target individual based on the nearness and sizes of neighbors. We applied a factorial regression analysis of survival and height growth of individuals as a function of hydrology and competition measured with a neighbor interference index.
Results/Conclusions
Hydrology is more important than interference by neighbors in affecting survival. The growth response of four out of ten species was unaffected by water level, neighbor interference, or their interaction. In contrast, the interaction of hydrology and competition was significant for four other species; trees located towards the dry end of the flooding gradient were more affected by neighborhood competition than trees at the wetter end of the gradient. Results have implications for the creation and rehabilitation of south Florida tree islands, whose number and area have declined in the past 20 years, as well as for forest restoration more generally.