The average bed shear stress and bed friction factor of samples with any roughness was derived from the head loss between upstream and downstream of a test section in an erosion tunnel. The method was validated in both hydraulically smooth (plexiglass; Reynolds number less than 25,000) and rough regimes (calibrated particles with known roughness). As a first step toward using this method on natural sediment, this method was tested with experimental mesocosms assembled from field collected materials (sieved sediments; diatoms). Bed shear stress measurement precision was high enough in the experiments to detect a positive significant relationship between bed friction factor and core roughness. The observed bed friction factor increase could be related to diatom growth but not to diatoms biomass. DOI: 10.1061/(ASCE)HY.1943-7900.0000627. (C) 2012 American Society of Civil Engineers.