- Ocean circulation is an important component in Earth's climate system. Predicting future climate and circulation changes requires an improved understanding of the past relationship between climate and ocean currents. The neodymium isotope composition (εNd) of water masses is frequently used as a quasi-conservative tracer to reconstruct ocean circulation. The current budget of Nd in the ocean cannot account for upwards of 95% of the Nd entering the ocean, hindering interpretations of the εNd tracer. This study aims to determine the magnitude of the sedimentary source of Nd to the global ocean, characterize the influence of the flux of Nd on the εNd distribution in the global ocean, and identify the diagenetic factors that control the magnitude of the flux. To address these goals, I compare pore fluids, overlying water column, and sediments from eight sites on the continental margin off Oregon and California (USA). These sites lie above, within, and below the Northeast Pacific's oxygen minimum zone and represent shelf-to slope settings at water depths between 80 and 3000 m. Concentrations of the rare earth elements (REEs), including Nd, in the pore fluids are up to two orders of magnitude higher than the REE concentrations in seawater. All pore fluid REE profiles exhibit shallow subsurface (2-10 cm) concentration maxima. Fractionation of the REEs occurs during mobilization and transport through the sediment column, indicated by changes in the rare earth patterns above and below the pore fluid REE concentration maximum. Based on the pore fluid concentration gradient, I calculate the benthic flux from the pore fluids to the ocean. These calculations show a flux of Nd to the ocean increasing from 3 pmol cm⁻² yr⁻¹ at our 200 m site to 32 pmol cm⁻² yr⁻¹ at our 3000 m site. No corresponding change in pore fluid phosphorous, iron, organic carbon, or silica among sites is observed. Additionally, the major mineralogy remains constant among sites. Extrapolating our flux estimates over the global ocean, I estimate a global benthic Nd flux between 18 × 10⁶ and 110 × 10⁶ mol annually. I conclude that the benthic flux of Nd to the ocean is the dominant source of Nd to the global ocean. The εNd of this flux ranges from -0.2 at our 200 m site, to -1.5 at our 1200 m site and, -1.8 at our 3000 m site. At our deepest (3000 m) site, the bottom water εNd (-2.3) is between the value expected for the water mass (-3.3) and the εNd of the flux (-1.8). The magnitude of this flux and the time a water mass is exposed to this flux determines the distribution of εNd. The degree of change in bottom water εNd depends on the difference between the initial εNd of the bottom water and the εNd of the flux. The benthic flux can determine the εNd of bottom water in a short enough time to create the heterogeneous εNd observed in the modern ocean. Additionally, the benthic flux provides a mechanism for the observed alterations in εNd in the deep North Pacific in the absence of Pacific deep-water formation.Based on εNd of the pore fluids, total sediment, and sediment leachates, I propose that the magnitude of the flux is a function of the authigenic coatings formed during sediment diagenesis. Because the pore fluid Nd represents less than 0.001% of the Nd in any given volume of the upper sediment column, changes in Nd must be driven by larger Nd reservoirs in the solid phases. The leachable Nd (acid leachable: 1 - 7 μg Nd g⁻¹ sediment; reducible 2 - 8 μg Nd g⁻¹ sediment) represents a large reactive Nd reservoir, accounting for ~50% of the Nd in any given volume. Because the total sediment digest and leachates are less radiogenic than pore water at our 200 m site, I infer that there are radiogenic trace mineral phases undergoing active exchange with the pore fluid reservoir. When present, this trace mineral disproportionally influences the εNd of the flux.The exchange of Nd between the authigenic coatings and the pore fluid in the upper sediment column means that the εNd of the true coatings will be the same as the pore fluids. However, leaching procedures result in a contaminated authigenic coating εNd signature, especially in the presence of reactive trace minerals. In regions with a large benthic flux, the authigenic coatings determine the εNd of the bottom water. Therefore, in these regions the coatings will resemble εNd of the bottom water. The εNd of the bottom water cannot drive the εNd of the coatings because the amount of Nd in the bottom water is quantitatively insignificant. Based on the findings that the benthic source is the dominant source of Nd to the ocean, the εNd of this flux can determine the εNd of the bottom water, and the εNd of the authigenic coating is not determined by the εNd of the bottom water, a reinterpretation of the εNd records is necessary. Because the εNd of bottom water is not conservative, the εNd record cannot be strictly interpreted as water mass mixing. The inverse relationship between ocean circulation speed and exposure time can result in differing εNd values with no change in circulation path. Additionally, the end member water mass εNd values are likely to have changed through time. Specifically, any shift in deep-water formation would result in the bottom water being exposed to a different εNd from the benthic flux. With reinterpretation that recognizes these influences, the εNd tracer may remain a useful proxy of past ocean circulation.