Estimating minimum thresholds of natural vegetation for the integrated management and protection of water quality in South African catchments

Locke, Kent

Estimating minimum thresholds of natural vegetation for the integrated management and protection of water quality in South African catchments

Thesis / Dissertation

2025

Publisher

University of Cape Town

Department

Environmental Humanities

Faculty

Faculty of Humanities

Abstract

Despite multi-level commitments to Integrated Water Resources Management (IWRM), many of South Africa's water quality problems are attributable to the negative impacts of anthropogenic land use on water quality. Academics and policymakers have warned that unless action is taken to improve water resources management through the implementation of coordinated, proactive, and data-driven strategies, the country faces a water crisis that will have severe socio-ecological consequences. As natural vegetation acts as a sink, thus protecting water resources from diffuse pollution, the preservation of an adequate amount within catchment areas is important. However, among several pertinent questions, it is not clear (1) how much natural vegetation cover is required, (2) at which scale(s) this would be most effective, (3) how natural vegetation should be classified, and (4) whether the fragmentation of natural vegetation is a significant factor. To answer these questions, regression analysis was used to model relationships between water quality (measured using a composite pollution index) and metrics of natural vegetation (estimated from national land cover maps) at multiple scales across a sample of sub-catchments located within South Africa's Berg-Olifants, Breede-Gouritz, and Mzimvubu-Tsitsikamma Water Management Areas. Across this sample, a statistically significant, nonlinear, and inverse relationship was found between proportions of natural vegetation cover and pollution levels. This relationship was strongest (1) when natural vegetation was defined as an aggregation of indigenous woody vegetation, wetlands, and forestry plantations, and (2) when measured across the whole catchment and within a 200 m riparian buffer zone. At both scales, however, fragmentation was not found to be significant. The models further indicated that approximately 82 to 90% natural vegetation cover was necessary at these scales to keep pollution scores within acceptable levels. Additional nonlinear thresholds estimated using breakpoint analysis also suggested that if proportions of natural vegetation fall below 45% (across the whole catchment) and 60% (within a 200 m riparian buffer zone) a dramatic increase in pollution levels can be expected. The study has direct relevance for IWRM in so far as these results demonstrate (1) the critical importance of preserving areas of natural vegetation for water quality management and (2) the possibility of providing actors with quantifiable and context-specific management targets which can inform multistakeholder decision-making processes at appropriate spatial scales.