Interactive effects of animal disturbance and elevation on vegetation of a tidal freshwater marsh

We studied interactions between animal disturbance (geese, carp, and muskrat) and elevation in a field experiment in tidal freshwater marshes of the Patuxent River, Maryland, United States. Vegetation changes were recorded in fenced and unfenced plots in high and low marsh community types for 2 yr using measurements of areal cover and within-plot frequency (which were averaged to create a dominance index), Leaf Area Index (LAI), and aboveground biomass. We related light environment to differences in vegetation using below-canopy measurements of Photosynthetically Active Radiation (PAR). In the low marsh, total cover of all species, cover of annual species, biomass, and LAI were significantly higher in plots fenced to exclude animals (exclosures) than in unfenced plots (fenced/unfenced total cover=76/40%, annual cover=45/10%, biomass=936/352 g m^?2, LAI=3.3/1.4). PAR was significantly lower in fenced than unfenced plots (fenced/unfenced=115/442 μmol s-1 m^?2). Despite the strong effect of fencing on biomass, species richness per plot (i.e., the number of species per plot, or species density) was not affected significantly by fencing in the low marsh. Most of the observed differences in cover, biomass, LAI, and PAR were due to variation in the abundance of the herbaceous annual speciesBidens laevis (dominance index fenced/unfenced=45/10%) andZizania aquatica (30/12%). In the high marsh community, fencing had only minor effects on plant community composition and did not significantly affect species richness, cover, biomass, PAR, or LAI. Our results show that animals can dramatically affect low marsh vegetation, primarily via physical disturbance or herbivory of shallowly rooted seedlings of annual species.