South Africa is an arid country and its growing population is putting freshwater resources under increasing pressure. Natural salinization of freshwater systems is being exacerbated by anthropogenic influences. The National Water Act (No. 36 of 1998) stipulates the need for an ecological Reserve, that quantity and quality of freshwater needed to protect freshwater ecosystems while allowing sustainable use of freshwater resources. Water guidelines do exist in the form of the South African Water Quality Guidelines (DWAF, 1996) and more recently, Jooste and Rossouw (2002) compiled benchmark values for water quality variables marking the boundaries between ecological health classes in the 4-category classification system. Predominantly international toxicity data were used to compile the guidelines and the benchmark values. In addition, there is a paucity of chronic toxicity data nationally and internationally. This thesis showed that it is statistically possible to derive protective chronic endpoints for salinity from acute toxicity data through extrapolation. The Acute to Chronic Ratio (ACR), Two-Step Linear Regression (LRA) and Multi-Factor Probit Analysis (MPA) extrapolation methods were investigated to derive chronic toxicity data from acute toxicity data. The authors of LRA and MPA recommend associating a time independent LCx value in the range of LC₀¸₀₁ to LC₁₀ with a Predicted No Effect Concentration (PNOEC). In addition to published methods, this thesis studied the possibility of equating a time independent LC₅₀ value and subjected to a safety factor of 5 (LRA LC₅₀/5), to the PNOEC. Extrapolated chronic toxicity data where the toxicants are NaCl and Na₂SO₄ were derived for indigenous South African macroinvertebrates. NaCl and Na₂SO₄ are salts associated with salinisation in South Africa. In addition, a chronic salinity toxicity test protocol for an indigenous South African aquatic macroinvertebrate was designed and chronic toxicity test were performed using NaCl and Na₂SO₄ as toxicants. The experimental chronic toxicity data produced were used to validate results from the acute to chronic extrapolation methods. Extrapolated chronic toxicity data were inputted into Species Sensitivity Distribution curves, and concentrations that were predicted to protect 95 % of species (PC95) were compared to the sub-lethality benchmarks proposed by Jooste and Rossouw (2002) for NaCl and Na₂SO₄. This study concluded that the LRA LC₅₀/5 extrapolation method is the most protective and accurate and proposed that LRA replace the ACR method in future guideline development for inorganic salts.

Although the use of pesticides is necessary to meet the socio-economic needs of many developing countries, especially in Africa, side effects of these bio-active chemicals have contributed to contaminating aquatic and terrestrial ecosystems. Environmental water quality degradation by pesticides interferes with ecosystem health and poses numerous risks to aquatic life. In South Africa, glyphosate-based herbicides are frequently used to control weeds and invading alien plants, but ultimately end up in freshwater ecosystems. However, there are no South African-based environmental water quality management strategies to regulate these bio-active chemicals. Therefore, this study sought to provide a sound scientific background for the environmental water quality management of glyphosate-based herbicides in South Africa, by conducting both laboratory and field investigations. In the laboratory investigations, aquatic ecotoxicological methods were used to evaluate responses of the freshwater aquatic shrimp Caridina nilotica exposed to Roundup® at different biological system scales, and the responses of multiple South African aquatic species exposed to Roundup® through species sensitivity distribution (SSD). In the field investigations, the effect of Kilo Max WSG on the physicochemical and biological conditions of three selected sites in the Swartkops River before and after a spray episode by Working for Water were evaluated through biomonitoring, using the South African Scoring System version 5 (SASS5) as a sampling protocol. Both Roundup® and Kilo Max WSG are glyphosate-based herbicides. All the data were subjected to relevant statistical analyses. Findings of this study revealed that Roundup® elicited responses at different biological system scales in C. nilotica, while SSD estimates were used to derive proposed water quality guidelines for glyphosate-based herbicides in South Africa. The biomonitoring revealed that using glyphosate-based herbicides to control water hyacinth within the Swartkops River had a negligible impact on the physicochemical and biological conditions. Based on these findings, a conceptual framework that can be used for the integrated environmental water quality management of glyphosate-based herbicides in South Africa was developed as part of integrated water resource management (IWRM). The combined data sets contribute to a sound scientific basis for the environmental water quality management of glyphosate-based herbicides in South Africa.