**3.1 Vegetation data**

A total of 18 sample plots were surveyed in the reference as well as restoration catchments of the four study sites. Plot sizes were fixed at 4 x 4 m to give a total surface area of 16 m² (Du Preez, 1991; Malan, 1998). All plant species growing in the sample plots were recorded and their cover abundance assessed using a modified Braun-Blanquet cover abundance scale (Brown, et al., 2005; Mueller-Dombois & Ellenberg, 1974; Westhoff & Van der Maarel, 1980).

Environmental data recorded in each sample plot included aspect, slope, exposure, the size of the rocks, latitude and longitude, altitude, locality, geology, the per cent of area covered by rock, topography, the degree of surface erosion (indicated as low moderate and severe), the degree of trampling, drainage, soil depth and notes on management and utilisation as well as total percentage canopy cover.

The habitat as well as floristic data was processed with the TURBOVEG database (Hennekens & Schaminée, 2001). A first approximation was done with the TWINSPAN (two-way indicator-species analysis) algorithm of Hill (1979). The JUICE program (Lubomír 2002) was used to generate a phytosociological table (Table 1), which was refined by Braun-Blanquet procedures.

#### **3.2 Biomass**

From a grazing and resilience point of view, production or yield is one of the most important measures in assessing an ecosystem (Carpenter et al., 2001). Biomass or standing crop usually refers to the weight of organisms present at one time (Pieper, 1978). Most estimates of plant biomass or standing crop include only that above the soil surfaces. This material is commonly available to large herbivores. Direct harvesting is considered the most reliable method of determining aboveground biomass (De Leeuw et al., 1991; Snyman & Fouché, 1993). For this study two exclosure plots of 25 m² were erected within each of the Khalong-la-Lithunya and Koti-Sephola peatland sites.

Biomass was determined in each of the exclosure and grazed plots as well in the other peatland systems by randomly placing 0.5 m2 quadrats in these areas and cutting the herbaceous material down to ground level (Figure 3). The grass and forb species were separated and dried a drying oven at 65°C for 48 hours (Cleaver, 2004).

The Kruskal-Wallis ANOVA by Ranks test (Welman et al., 2007) was used to compare the biomass (kilogram Dry Matter/hectare)(kg DM/ha) of the plots by grazing type while the

Impact of Domestic Animals on Ecosystem Integrity of Lesotho High Altitude Peatlands 255

an average height of 4 cm while the grazed plots had an average height of 2 cm. No difference

Table 1. Phytosociological table of the major high-altitude wetlands of Lesotho and Platberg.

in species composition was found between the exclosure and grazed plots (Table 1).

the Mann-Whitney U Test (Welman et al., 2007) was used on the ranked data to determine where the pairwise differences lie.

Fig. 3. All rooted herbaceous plants were harvested on a species basis in quadrats (0.5m2), randomly placed in the grazed and exclosure plots in the peatlands.
