Recent advances in streamflow discharge measurement instrumentation have improved upon or following replacements of conventional devices and methodologies. Among the advanced instruments, surface velocity radar (SVR) has been increasingly used for flood measurements and normal flow discharge, thus enabling noncontact measurements of surface velocities by using microwaves. It is effective even when direct access to rivers is difficult, such as during the flood season. Unfortunately, systematic and comprehensive evaluation of surface velocity measurement accuracy in conditions in situ is rarely conducted. The microwave footprint on the water surface, as emitted from the SVR, depends on the monitoring configurations (e.g., tilt and yaw angles) and can affect the accuracy of velocity measurements. In this study, we selected an SVR with a rectangular front shape and evaluated its measurement accuracy in a real-scale river experiment facility wherein surface velocity was validated against concurrently measured, large-scale particle image and acoustic Doppler velocimetry. Comparative results showed that the spatial autocorrelation of the velocity distribution inside the SVR footprint was crucial such that the measurement configuration for low-tilting and high-yawing angles should be avoided especially for locations near the channel edge with substantially reduced velocity.