Pickens, B.A.; Taylor, J.C.; Finkbeiner, M.; Hansen, D., and Turner, L., 2021. Modeling sand shoals on the U.S. Atlantic shelf: Moving beyond a site-by-site approach. Journal of Coastal Research, 37(2), 227–237. Coconut Creek (Florida), ISSN 0749-0208.The demand for offshore marine sands has escalated worldwide as sediments are needed for increasingly frequent beach renourishment and barrier island restoration. Sand shoals are often used as a source for dredging material because of the high volume of sand per unit area. Yet, investigations of shoals are typically conducted on a site-by-site basis, and a broader understanding of shoal availability is needed for strategic decision-making, including the mitigation of ocean use conflicts. Here, the primary objective was to model shoal distribution across the U.S. Atlantic shelf, including the Gulf of Mexico. Publicly available bathymetry data were obtained at a relatively coarse 90-m resolution. Variables of depth, standard deviation of depth, slope, bathymetric position index, and distance to shoreline were used as predictors to identify shoals. Unsupervised classifications of the seafloor were conducted to distinguish shoals and swales. Classification accuracy was assessed with validation databases of identified sand resources and named shoals compared to random locations; a visual assessment was also conducted. Shoals were further characterized by their origin. The classifications showed shoals and swales differed from the seafloor. Shoals were more shallow, had higher slope, a higher standard deviation of depth, were closer to the shoreline, and had a more positive bathymetric position index. Shoals were classified on 4.7% of the U.S. Atlantic shelf, and validation showed a percent agreement of 65–93%. Classified shoals visually coincided with the shape and extent of known sand resources. Shoals were characterized as cape-associated, bedform, isolated shelf, or uncharacterized. For the continental shelf, multivariate predictors represented the heterogeneous sloping substrates and the flat, high relief crests of sand shoals. The ability to classify shoals with 90-m resolution bathymetry data in the U.S. Atlantic reveals the methodology may be applicable to identify sand shoals elsewhere in the world with currently available data.