Composition and diversity of Mopaneveld herbaceous vegetation : an exclosure experiment

Abstract

The herbaceous vegetation layer of the semi-arid Mopaneveld savanna possess a distinctive homogeneity and is highly dynamic in its response to disturbances such as herbivory. Understanding how this ecosystem reacts to selective herbivore exclusion, provides valuable information for conservation management. Dynamic responses of the herbaceous vegetation in the past have always been tested on either plant species or functional groups. Very little is known of the herbaceous dynamics if a direct comparison is made of simultaneously tested responses on species and functional level and whether one approach will be more sensitive to detect changes in the herbaceous community than the other. The Letaba research exclosures in the Kruger National Park (KNP), provided opportunities to investigate how long-term herbivore exclusion effects composition, richness and diversity of the herbaceous layer in a Mopaneveld savanna at both species and functional group levels. The study aimed to analyse the herbaceous response patterns while specifically focussing on (a) the quantification of changes in herbaceous plant community structure, (b) how species composition, richness and diversity respond at both species as well as functional group level and (c) assessment of community stability as given by the variability of assessed parameters and calculated measures. The three herbivore exclusion treatments of the Letaba exclosures consisted of (1) a fully fenced area, excluding all herbivores larger than a hare, (2) a partially fenced area, excluding only elephants and giraffes and (3) an open, unfenced control area. Field data collection entailed sampling herbaceous vegetation in two 1 m2 circular sub-plots in the eastern and western corners of each of the 151 fixed plots (data of subplots were then pooled for a better species representation per plot). Biomass data for each plot were sampled using a Disc Pasture Meter (DPM). Literature-based soft functional traits representing adaptations to herbivory in a semi-arid system were used to classify all sampled species into functional groups. Repeated measures ANOVA was used to test for the significant effect of herbivore exclusion over time. Non-Metric Dimensional Scaling was used for the species level data and one-way ANOVA for the functional group data to test for composition changes. Cluster analysis and Principal Component Analysis were used to identify refined plant functional types. After nine years of herbivore exclusion, few significant response patterns were revealed at both species and functional levels. At the species level, no significant herbivore exclusion effect could be detected over time for either composition, richness or diversity. Herbivore exclusion led to a significant increase in grass biomass (above-ground phytomass). The exclusion of large herbivores (i.e. elephant and giraffe), promoted herbaceous community evenness. At the functional group level, perennial groups dominated. Since the exclusion of both all herbivores and just large herbivores resulted in increased dominance by perennials, it was concluded that this change was due to the effects of the large herbivores only and that large herbivores are required to maintain a more evenly distributed composition of functional groups. In terms of diversity, differences between treatments were detected at the start of the experiment, but significant time and treatment effects indicated that the absence of only large herbivores promoted a higher and more evenly distributed functional group diversity. Despite dynamic fluctuations in the herbaceous community structure that could have occurred between 2003 and 2012, the herbaceous community did not change significantly on species or functional group level, which may suggest high resilience to the environmental disturbances present at the study site.