Stream synoptic sampling studies that include flow estimates derived from the stream tracer dilution method are now commonly performed to identify sources and processes controlling solute transport to streams. However, a limitation of this mass-loading approach is its inability to identify the side of the stream on which a source is located in the common case where loading is largely from groundwater discharge. Such bank-specific loading information can be particularly valuable at mining-affected sites where both mining-related and natural background metal sources are often closely intermingled. Here we combine the stream mass-loading approach with data from pairs of shallow hand-installed streambank wells located on opposite sides of a gaining metal-impacted headwater stream to estimate bank-specific metal loading rates. Study results successfully identify a right-bank zone in the upper half of the primary study reach as the dominant source of loading, where groundwater discharge is elevated in metals due to natural weathering of sulfide-rich bedrock. A left-bank zone in the lower half of the primary study reach was also identified as a secondary, yet still substantial, loading source, where groundwater is likely impacted by portal discharge and/or waste piles associated with an adjacent abandoned mine. Determining the dominance of the left-bank mining-related source compared to other potential right-bank natural sulfide weathering sources would not have been possible without the streambank wells. Streambank wells also enabled collection of dissolved gas samples for groundwater dating. Computed piston-flow ³H/³He groundwater ages show little to no correlation with either metal concentrations or loading rates, suggesting that groundwater residence time and flow path variations exert little influence on groundwater discharge chemistry compared to variations in bedrock and soil composition. This study thus provides proof of concept that the proposed method of combining streambank well data with stream tracer injection and synoptic sampling can provide useful information on bank-specific mass loading rates and sources of contaminants affecting stream water quality.