Lakes and rivers are vital sources of potable water in many countries, but taste and odour (T&O) compounds such as geosmin and 2-methylisoborneol (2-MIB) can impact water quality aesthetics and require costly treatment processes. In this study, we analysed more than 10 years of historical monitoring data collected by a water authority for a subtropical reservoir and its downstream river, which together form the primary drinking water supply for Brisbane, Australia (population 2.7 million). To identify sources of T&O in the river and potential zones of removal, we assessed changes in T&O concentrations along 62 km of the Brisbane River, which transports water from the reservoir to the water treatment plant offtake. Geosmin and 2-MIB exhibited distinctly different spatial patterns within the system. 2-MIB in the river primarily originated from the upstream reservoir, with concentrations generally declining downstream. There were no consistent patterns between 2-MIB concentrations and the abundance of cyanobacterial species which have been linked to production of this compound in the reservoir. Possibly, production by non-photosynthetic bacteria in the reservoir’s bottom waters may contribute to 2-MIB levels. In contrast, the upstream reservoir was rarely a source of geosmin to the river since geosmin increased in the upper reaches of the river at various times, when there were no tributary inputs. This suggests an in-river source of production of geosmin. Targeted microbial sampling revealed that geosmin concentrations were correlated with the abundance of several actinobacteria species in the river water column. Concentrations of both 2-MIB and geosmin typically decreased downstream, particularly along the lower river reaches. As a result, despite upstream sources from both the reservoir and river, concentrations at the downstream water treatment plant were generally low (<10 ng L⁻¹). The removal of geosmin and 2-MIB by natural degradation processes during transit has significantly reduced the T&O concentrations arriving at the water treatment plant, both during major T&O events in the upstream reservoir and under background conditions. Achieving an equivalent level of removal through treatment plant upgrades (e.g., powdered activated carbon filtration) would require capital investments of several million dollars and incur significant ongoing operating costs. These findings suggest that both cyanobacteria and non-photosynthetic bacteria may contribute to 2-MIB and geosmin concentrations in this system, with each compound being sourced from different habitat zones. Different strategies may therefore be needed to predict and manage each compound in this water supply system.