Abstract Understanding groundwater ages within an aquifer system has the potential to better constrain estimates of groundwater recharge and flow rates, and therefore increase the reliability of groundwater models. Groundwater ages are generally interpreted from field-observed environmental tracer concentrations, but in many cases in which multiple groundwater age tracers have been analyzed simultaneously the results show significant disparities among tracer-specific estimated ages. The disparities are generally attributed to physical mixing between waters of different ages. However, especially in the geochemical literature environmental tracer concentrations are often analyzed with simplistic models in which the degree of the simulated mixing might be considered unrealistic for natural heterogeneous geologic media. In this study we use numerical experiments to examine under which physical conditions measured concentrations of selected environmental tracers (CFC-12, 39Ar, 14C) may return discrepant ages. Our model simulations suggest that matrix diffusion has the greatest potential to cause mixing of different-aged water and to generate age biases between tracers. The multi-tracer simulations also suggest that there is a limit to the magnitude of the discrepancies that can be attributed to physical processes. When comparing data collected from the Pilbara region of Western Australia with our numerical modeling studies, it was found that a dual-domain mass transfer model was required to explain the field-observed age-discrepancies.