Re-evaluation of groundwater residence time using a combined 3H/3He, 14C and 222Rn approach

Abstract

<p>The TMG aquifer is one of the largest aquifer systems in South Africa and is currently targeted as a potential source of potable water for the City of Cape Town (CoCT) which recently experienced a period of extreme water stress. Groundwater in the TMG aquifer typically has very low total dissolved salts, on the order of 50 mg/L of less, making it challenging to constrain the groundwater residence time. However, residence time is a key parameter to provide proper constraints on turnover time of groundwater in the aquifer system before large-scale abstraction takes place, in order to evaluate the sustainability of the resource. This study used the <sup>3</sup>H/<sup>3</sup>He system to date modern water (<100 years) and <sup>14</sup>C to date older groundwater (>500 years). Groundwater residence times were determined for the TMG aquifer and five associated aquifer systems in the Western Cape of South Africa, namely the alluvial, Witteberg, Bokkeveld, Cape Granite Suite (CGS) and Malmesbury aquifers. Good correlation between <sup>3</sup>H/<sup>3</sup>He and <sup>14</sup>C ages indicate relatively short residence times for the alluvial and TMG aquifers whereas groundwater from the Witteberg, Bokkeveld, CGS and Malmesbury aquifers indicate mixing of older water bodies with modern recharge resulting in distinctly different ages derived from the two dating systems. In an attempt to better constrain the mixing relationship with modern precipitation, <sup>222</sup>Rn was used to assess the interaction between precipitation and groundwater after rainfall events. The basis for this approach comes from the assumption that precipitation has little <sup>222</sup>Rn in it, with groundwater <sup>222</sup>Rn derived from interaction with the groundwater host rocks. This should result in groundwater <sup>222</sup>Rn activity being diluted through recharge with precipitation. However, since the half-life of <sup>222</sup>Rn is only 3.82 days, <sup>222</sup>Rn activities should respond rapidly to recharge, and should also recover rapidly from this recharge. Three behavioural characteristics were established; (1) groundwaters where the <sup>14</sup>C activity was of ≥ 100 pMC (TMG and alluvial aquifers), and where an immediate dilution in radon’s activity was recorded due to direct recharge. (2) groundwaters where the <sup>14</sup>C activity was 80% – 90% pMC (Malmesbury aquifer) where a delayed response in the dilution of radon’s activity was recorded; and (3) groundwaters where the <sup>14</sup>C activity was ≤ 70% and radon activities were stable indicating little or no recharge. <sup>222</sup>Rn thus proved an important mechanism for evaluating the validity of residence times derived from both <sup>3</sup>H/<sup>3</sup>He and <sup>14</sup>C.</p>