We investigated the role of increasingly well-constrained geologic structures in the subsurface (i.e., subsurface architecture) in predicting streambed flux and hyporheic residence time distribution (RTD) for a headwater stream. Five subsurface realizations with increasingly resolved lithological boundaries were simulated in which model geometries were based on increasing information about flow and transport using soil and geologic maps, surface observations, probing to depth to refusal, seismic refraction, electrical resistivity (ER) imaging of subsurface architecture, and time-lapse ER imaging during a solute tracer study. Particle tracking was used to generate RTDs for each model run. We demonstrate how improved characterization of complex lithological boundaries and calibration of porosity and hydraulic conductivity affect model prediction of hyporheic flow and transport. Models using hydraulic conductivity calibrated using transient ER data yield estimates of streambed flux that are three orders of magnitude larger than uncalibrated models using estimated values for hydraulic conductivity based on values published for nearby hillslopes (10−4 vs. 10−7 m2/s, respectively). Median residence times for uncalibrated and calibrated models are 103 and 100 h, respectively. Increasingly well-resolved subsurface architectures yield wider hyporheic RTDs, indicative of more complex hyporheic flowpath networks and potentially important to biogeochemical cycling. The use of ER imaging to monitor solute tracers informs subsurface structure not apparent from other techniques, and helps to define transport properties of the subsurface (i.e., hydraulic conductivity). Results of this study demonstrate the value of geophysical measurements to more realistically simulate flow and transport along hyporheic flowpaths.