Accurate stream topography measurement is important for many ecological applications such as hydraulic modeling and habitat characterization. Habitat complexity measures are often made using visual approximations or total station (TS) surveying that can be subjective and have limited spatial resolution. Terrestrial laser scanning (TLS) can measure topography at a high resolution and accuracy. Two methods, TS surveying and TLS, were compared for measuring complex topography in a boulder-dominated 100 m forested reach of the Staunton River in Shenandoah National Park, Virginia. The mean absolute difference between the two datasets was 0.11 m with 82.3 percent of the TS data within _0.1 m of TLS. The TLS dataset was processed to remove vegetation and create a 2 cm digital elevation model (DEM). An algorithm was developed for delineating rocks within the stream channel from the DEM. A common ecological metric based on the structural complexity of the stream, percent in-stream rock cover, was calculated from the TLS dataset, and the results were compared to estimates from traditional methods. This application illustrates the potential of TLS to quantify habitat complexity measures in an automated, unbiased manner.