The role of termite mounds in vegetation transition to woody plants in grassy ecosystems in KwaZulu-Natal, South Africa.

Abstract

Woody plants encroachment, the increase in woody plant cover and density in grassy ecosystems, poses a significant threat to biodiversity and ecosystem functioning across southern Africa and globally. While drivers such as elevated atmospheric CO₂, climate change, land use and land management practices contribute to this phenomenon, landscape-scale processes such as termite activity plays an important role. Termites, through mound-building and soil modification, create nutrient-rich and stable microhabitats that facilitate woody plant establishment and the formation of bush clumps. This study investigated the role of termite mounds in promoting bush clump development and driving woody plant community dynamics in Cumberland Nature Reserve and Uitkomst Conservation Area in KwaZulu-Natal, South Africa. Specifically, the study aimed: (1) to establish the growth rate and expansion of bush clumps facilitated by termite mounds; (2) to assess the influence of termite mounds on woody species richness, diversity, and composition, including the role of bird-mediated seed dispersal in amplifying woody plant establishment; (3) to explore changes in woody species richness, diversity, and composition in relation to termite mound size, and assess the relationship between spatial distribution of termite mounds and woody vegetation cover; and (4) to investigate the patterns and processes of woody plant succession associated with bush clump development on termite mounds. Using a time series of aerial photographs (1964, 1983, 2003, and 2024), the development and expansion of bush clumps were studied in Uitkomst. Bush clumps were consistently found to initiate on termite mounds, with the earliest clumps present before 1964. Clumps recorded before 1964 were also the largest by 2024, suggesting that clump size reflects successional age. From 1964 to 2024, bush clump area increased significantly, tripling in size. Growth rates of bush clump area varied with land-use type: grassland clumps expanded at 4.7 m² year⁻¹ from 1964 to 2003, whereas in the cropland area, clumps grew at only 0.15 m² year⁻¹ during the same period. However, following cropland abandonment post-2003, clumps in these areas expanded rapidly (~17.51 m² year⁻¹), with a tenfold increase in area between 2003 and 2024. Vegetation surveys on 62 termite mounds and 71 matrix plots across Cumberland and Uitkomst recorded 2432 plants representing 54 woody species. Termite mounds supported significantly greater woody plant density and diversity compared to the matrix: 72% of species were moundexclusive, and mound plots exhibited higher sapling and mature tree densities, stem density, species richness, and Shannon diversity index. Notably, matrix vegetation was dominated by Vachellia sieberiana, while termite mound assemblages were dominated by Lantana camara, Rubus ulmifolium, Cussonia spicata, and Searsia pyroides. Lantana camara, C. spicata, and S. pyroides were significant termite mound indicator species (indicator value < 60%), whereas none were identified for matrix assemblages. Bird-dispersed species with drupe-type fruits were more common on mounds, suggesting avian dispersal influences species composition. To test whether the size and spatial distribution of termite mounds influenced the woody plant composition and distribution of bush clumps, woody vegetation assemblages on bush clumps occurring on termite mounds of varying sizes were studied. In addition, termite mound density, bush clump density, and the number of termite mound–facilitated clumps were quantified within 41 open-area plots at both Cumberland and Uitkomst to assess spatial relationships between mound occurrence and bush clump formation. Termite mound density ranged from 4.7 to 5.4 mounds ha⁻¹, with bare mound densities at between 1.2 and 2.4 mounds ha⁻¹ and vegetated mounds at between 3 and 3.4 mounds ha⁻¹. Mound surface area varied from 1.46 m² to 111.12 m². Bush clump density was significantly and positively associated with termite mound density (R² = 0.512, F(3, 37) = 12.92, P < 0.0001), indicating that greater mound densities correspond to greater woody clump densities. Species richness of woody plants increased with mound surface area (R² = 0.362, F(3, 58) = 10.96, P < 0.0001), and the proportion of termite mound indicator species was significantly higher on large mounds than on medium and small mounds. Woody plant density and sapling density were ~3 times higher on large mounds, and mature tree density was ~2 times higher than on smaller mounds. Stem density was ~4 times higher on large mounds. A space-for-time substitution approach was employed to investigate the developmental processes of bush clumps. Sixty-two bush clumps representing different successional stages were sampled to identify potential founder species and to examine changes in species composition as bush clumps matured and increased in size. Bush clump area ranged from 4 to 450 m² (mean ± SE = 102.14 ± 10.67 m²). Each clump contained between 5 and 128 woody plants, averaging 28 ± 3 individuals. Species richness ranged from 2 to 20 species per clump (5.6 ± 0.41), and Shannon diversity index averaged 1.27 ± 0.07. Dominant species included L. camara, R. ulmifolius, Lippia javanica, C. spicata, and S. pyroides. Evidence indicated that bush clumps in both Cumberland and Uitkomst were simultaneously initiated by multiple species, with S. pyroides and V. sieberiana identified as the founder species in approximately 50% of the bush clumps. Species composition differed significantly between small and large bush clumps (PERMANOVA R² = 0.19, P = 0.001), and compositional similarity (MorisitaHorn index = 0.76 ± 0.03) between saplings and mature trees supports a deterministic successional trajectory. A lognormal regression revealed that tree species richness, abundance, Shannon diversity, woody basal area, and stem diameter all increased significantly with clump size, indicating that bush clump expansion is accompanied by increased species richness, diversity, and structural complexity, consistent with successional development driven by founder establishment and ongoing recruitment. The proportion of representatives of founder species declined with clump size, while the proportion of alien woody species increased. However, community-weighted means for forest specialisation and species habitat breadth did not vary significantly with clump size.