Materials and Methods

Description of the Study Areas

The study was conducted in the Jigjiga field site (9.356784 N, 42.795519W) of the Somali Regional State of Ethiopia (Figure 1), which is part of a pastoral ecosystem. This area experiences two rainy seasons: the main season occurs from July to September, and the shorter one is from March to April. The average annual rainfall in the area is 660 mm (Gebremedhn et al. , 2022) and the soils are mainly sandy loams (Alemie and Gebremedhin, 2019). The temperature is generally high throughout the year, with mean minimum values around 20 °C and mean maximum values around 35 °C (Gezahegn, 2006). The natural vegetation of the area is Acacia wooded grasslands, with Chrysopogon aucheri and Eragrostis spp, being the most dominant grass species and Acacia ethbaica, Acacia busse, and Vachellia nitlotica . are being the most dominant woody species of the study area (Hailu 2017). The land-use system in the area is primarily pastoral, with the local community being nomadic and relying heavily on livestock grazing for their livelihoods.

Site selection and sampling design

Before selecting the study plots and sampling techniques, a preliminary survey was conducted with local natural resource management experts and elders who had extensive knowledge of grazing management practices in the study area. Based on their input, three traditional rangeland management practices were chosen for this study: enclosures, communal open grazing, and browsing land. (i) Enclosures are used for hay production, which is cut and carried to the livestock when there is a feed shortage for grazing in the open communal grazing areas. (ii) Communal open grazing areas are characterized by open grass vegetation with scattered trees, and are used for extensive livestock grazing throughout the year. (iii) Browsing land, also known as “bay land” and is dominated by bush vegetation and is used for camel and goat browsing. The communal open grazing area represents the most common land-use system in the Somali rangelands. For a detail description of the management systems of the field sites see Gebremedhn et al., (2022).
The transect survey method was used for sampling vegetation attributes across the three different management practices. Specifically, nine square plots of 400 m2 each were established at an interval of 5 km for grazing land from Harishin to Kebri Beyah rangelands of the Jigjiga zone (Figure 1). Similarly, nine 400 m2 square plots were laid at an interval of 1 km for browsing land from Awebere rangelands of the same zone (Figure 1). Whereas for sampling enclosure sites, three 400 m2plots were randomly placed within each of the three private enclosures aged 20 to 30 years, for a total of nine plots (three plots within every three enclosures). This methodology allows for representative sampling of vegetation attributes within different management practices and was chosen as it can help to identify differences and similarities between them.

Vegetation Sampling

To quantify the structure of woody vegetation, we measured tree/shrub densities, canopy diameters, canopy heights, and stem heights of identified woody species in each 400 m2 plot. Canopy cover was calculated using the average of the two longest canopy diameters perpendicular to each other and parallel to the ground, following the method of Greig- Smith (1983). Stem height was measured as the total height of the plant stems from the ground level to the highest foliage. For species with multiple stems, each stem was measured separately, and the average was taken. Height measurements and canopy lengths and widths were conducted for the whole plant by measuring multiple stems as if it was one tree. To estimate woody aboveground biomass (AGB) in a non-destructive way, we used biomass regression equations (allometric equations) developed by Hasen-Yusuf et al. (2013).
For sampling herbs vegetation, we placed five sub-quadrats of 1 m2 in a zigzag pattern (i.e., four at all corners and one at the middle position of each 400 m2 plot, (as shown in Figure 2), making a total of 135 plots. From each of the 1 m2 quadrats, we collected samples of species, richness, composition and biomass. We determined species richness as the sum of all plant species present in the 1 m2 quadrats. The nomenclature of the plant species followed the Flora of Ethiopia (Hedberg and Edwards 1995). We estimated herbaceous species frequency by dividing the total number of quadrats in which the species occurred by the total number of quadrats studied in the 1 m2quadrats. The recorded species were categorized into three desirability classes based on their preference for grazing by livestock animals, using local ecological knowledge derived from herders and documented literature (Jerry et al. 1989). Additionally, we identified all herbaceous vegetation within the plots as either grass or non-grass species (forbs) following Behnke (1986). We estimated aboveground herb biomass by harvesting live and dead material at ground level. We weighed the harvested samples in the field to obtain fresh weight. Thirty percent of the harvested samples from each quadrat were placed in a paper bag for later dry matter analysis. This harvested biomass was dried in an oven at 105°C for 48 hours and then weighed to obtain the dry matter. All measurements were made from September to December for all study plots when the vegetation was at its peak flowering stage.

Data Analysis

All statistical analyses were carried out using R Statistical Software version 4.1.1. (R Core Team 2020). To determine the impact of traditional grazing management practices on herbaceous species composition, we used Canonical Correspondence Analysis (CCA) test on the frequency of herbaceous species present in the 1 m2plots, and the “anova.cca” function in the vegan package in R Statistical Software (Ter, 1986). CCA is a multivariate method that examines the relationship between species and their environment (Aminet al. , 2023).The ordination diagram generated from CCA describes the differential habitat preferences of taxa based on gradients. To determine the impact of traditional grazing management practices on woody species composition, we used the Analysis of Similarities (ANOSIM) test on the number of each woody species counted in the 400 m2 plots, and the “anosim” function in the vegan library R Statistical package. ANOSIM is a non-parametric test that compares groups of samples based on any distance measure (Clarke and Ainsworth, 1993). We also performed Analysis of Variance (ANOVA) tests on species richness, biomass, woody density, and canopy cover using the “aov” function to determine the effects of traditional grazing management practices. The student–Newman–Keuls post hoc test for differences in means, performed using the SNK.test function under theagricolae package (version 1.4.0), was used to compute significant differences among management practices.