Distance-based variations in tree seedling species regeneration in traditional agroforestry of the vhembe biosphere reserve, South Africa

Abstract

Seedling regeneration and β-diversity are critical for maintaining tree structure, biodiversity, and ecosystem functioning in traditional agroforestry landscapes (TAL). This study assessed regeneration seedling density (<5 cm DBH), species richness, and β-diversity across a spatial gradient in the Vhembe Biosphere Reserve, South Africa. Sampling was conducted in 675 subplots nested within 135 plots across three distance classes: immediate (0–1 km), intermediate (1–5 km), and far (5–10 km). A total of 4693 seedlings representing 117 species were recorded. Diversity patterns were analysed using SIMPER, PERMDISP, PERMANOVA, and distance-based linear modelling (DISTLM). Significant differences ( p < 0.05) in regeneration attributes were observed across distance classes. Seedling density increased with distance from settlements, with the highest values at far distances, while species richness and β-diversity peaked at intermediate distances. Species composition differed significantly among distance levels, with the greatest dissimilarity between the immediate and far zones. A few resilient and multipurpose species, including Tabernaemontana elegans, Grewia microthyrsa, and Combretum molle, were consistently abundant across all distance levels. DISTLM identified distance, elevation, slope gradient, and grazing as key drivers of regeneration patterns, with distance being the most influential. Higher disturbance near settlements, associated with harvesting, grazing, and land-use pressure, reduced regeneration density and diversity. The findings show that although TAL systems support substantial biodiversity and reflect community-based conservation of indigenous trees, regeneration is constrained by anthropogenic disturbances. Sustainable management, including reduced disturbance, regulated grazing, and protection of natural regeneration, is essential for long-term sustainability.