Vegetation community dynamics of a tropical semi-arid system following experimental removals of an exotic grass

Background/Question/Methods

Exotic species are among the primary threats to the conservation and management of biodiversity. Grasses comprise a group of plants that have been moved actively by humans around the globe for centuries and are well-known invaders of many ecosystems. Tropical dry forests are one of the most endangered and least conserved biomes in the world, and represented areas where grass invasions are common events. In this study we asked whether or not the presence of the African grass Megathyrsus maximus on Mona Island Reserve, a Caribbean dry forest, was associated with changes in abundance and composition of native vegetation. To address this question we performed vegetation surveys (from May 2009 to December of 2010) of plant species richness and abundance in plots established in areas invaded and non-invaded by the grass, as well as in areas where grass was completely or partially (trimming) removed.  We compared temporal changes in vegetation among treatments using the successional vector overlay of the non-metric-multidimensional scaling ordination (NMS).  Additionally, we used regression analysis to test for associations between changes in species richness and plant abundance with changes in environmental conditions.

Results/Conclusions

All sites considered were highly heterogeneous in vegetation composition.  Areas without grass had more vegetation than areas with the grass but species richness and the types of species found were very similar.  Removing the grass (completely and partially) did not increase species richness or abundance but plots where these two treatments were applied had similar temporal dynamics to plots dominated by the African grass. Species richness was dynamic but only positively associated with precipitation in unmanipulated plots with and without the grass. Plant abundance changed through time but changes were not significantly associated with precipitation in any of the plots. Observed temporal changes in community composition suggest that non-invaded plots are more stable communities than plots invaded by the grass and plots where the grass was removed.