Changes in streamflow contributions with increasing spatial scale in Thukela basin, South Africa

Sustainable management of river basins requires precise understanding of the origin and variability of water fluxes. Water samples were collected in Thukela Basin (30,000 km2), South Africa, over the 2012 rainy season, from fifteen 1 m2 runoff microplots (for OF), a 5-m deep piezometer (SW) and 20-m deep borehole (GW), in the basin headwater and nested catchment outlets (microcatchment, 0.23 km2; subcatchment, 1.20 km2; catchment, 9.75 km2; sub-basin, 253 km2). The water samples were analysed for Sodium (Na) and Silica (Si) concentrations using an inductively coupled-plasma emission spectrophotometry. End Member Mixing Analysis (EMMA), with Na and Si as tracers, was then used to quantify the water compartment contributions to river flow. The results showed a general decrease of unit-area runoff in downslope direction from 5.7 to 1.2 L m−2 day−1 at microplot and microcatchment level, respectively, to 1.4 L m−2 day−1 at the basin outlet. OF contributions averaged 61% at microcatchment, 79% at subcatchment, 40% at catchment, 78% at sub-basin and 67% at the basin outlet, which corresponded to 0.82, 0.26, 5 × 10−5, 2 × 10−3 and 9 × 10−5 L m−2 day−1, respectively. The respective SW contributions were 39% (0.38 L m−2 day−1), 18% (0.10 L m−2 day−1), 49% (5 × 10−5 L m−2 day−1), 15% (4 × 10−4 L m−2 day−1) and 33% (5 × 10−5 L m−2 day−1). GW contributions were much lower at all spatial scales, but showed a general increase with increasing contributing surface area from microcatchment to sub-basin outlet followed by a decrease to the basin outlet. The end-member contributions showed large spatial variations, hence longer-term research integrating more observation points is recommended to generate adequate data for development of prediction models for this important river basin. More research linking carbon, nutrient and pollutant fluxes to water dynamics is also recommended.