Downward fluxes of organically rich biodeposits under suspended mussel lines can cause benthic impacts such as changes in benthic community structure or microbial mat production. Quantifying sediment erosion in these coastal ecosystems is important for understanding how fluxes of organic matter and mussel biodeposits contribute to benthic–pelagic coupling. Critical shear velocity (u*crit), erosion rates and particle size distributions of resuspended sediment were measured at four stations distributed along a transect perpendicular to a mussel farm in Lagune de la Grande Entrée, Îles-de-la-Madeleine (Quebec, Canada). Stations were selected underneath the outer-most mussel line (0 m) and at distances of 15, 30 m and at a reference station (500 m) further along the transect. Shear velocity was measured using a calibrated portable Particle Erosion Simulator, also referred to as the BEAST (Benthic Environmental Assessment Sediment Tool). Undisturbed sediment cores obtained by divers were exposed to shear stress to compare differences between stations. Erosion sequences indicated no significant differences in u*crit between stations, but there were significant differences in erosion rates beneath mussel lines compared to other stations. Erosion rates were the highest in cores from beneath mussel lines, but paradoxically had the lowest u*crit. Mean erosion rates at u*crit varied between 25 and 47 g m− 2 min− 1 and critical erosion thresholds varied between 1.58 and 1.73 cm s− 1, which compare with intensive mussel culture sites elsewhere in eastern Canada. Significant differences existed in biotic and abiotic properties of sediments which could explain variation in maximum erosion rates within and between stations. Particle sizes measured by videography of resuspended sediment at different shear velocities ranged from 0.2 to 3.0 mm. Quantifying sediment erosion from intact marine sediments helps to improve our mechanistic understanding of these processes, and the BEAST further contributes to predictive capability in benthic–pelagic coupling modeling.