Temporal and spatial differences between predicted and measured organic carbon in South Atlantic sediments: Constraints to organic facies modelling

The evaluation of total organic carbon (TOC) content from wells is a routine practice for assessing source rock potential, and empirical relationships to estimate TOC content have been proposed and used in organic facies modelling. However, no critical evaluations address how well the available TOC-predictive equations estimate the TOC content in recent and past sediments of the South Atlantic. In order to investigate the performance of the available TOC-predictive equations in this study, we compiled a large South Atlantic dataset composed of new and published modern TOC, dry bulk density, and sedimentation rate with the associated satellite-derived net primary productivity. We repeat this exercise for another time interval using the same input parameters but estimating net primary productivity using transfer functions. Our results indicate that all equations overestimate the modern TOC distribution, mainly at the continental margins and at the edges of the subtropical gyre. However, we must point out that this is partly due to our primary productivity values used as inputs, which exhibit higher values than the ones used in the calibration datasets of the TOC-predictive equations. In addition, the equations perform better, in terms of correlation, in the eastern compared to the western South Atlantic. We observe that in the west, the equation of Muller and Suess (1979) exhibited the highest correlation between predicted and measured TOC values, while in the east, the Definition equation using the carbon flux equation of Antia et al. (2001) and the burial efficiency equation of Betts and Holland (1991) showed the best results. In addition, we indicate that lateral transport of organic carbon at the margins might be influencing the correlation negatively. Overestimations the TOC content were also apparent for the downcore data, although with reduced magnitude, and correlation between predicted and measured TOC values was lower compared to the modern dataset. This can be explained by more sparse data in the glacial dataset and a strong lateral transport during this interval of low sea level. The reasons behind the different performances of equations are explained by the way they incorporate the input parameters into their formulations, which is especially relevant for primary productivity, sedimentation rates, and the inclusion of processes such as the carbon flux. We highlight that carbon flux equations should also be used with caution in very shallow and high productive systems, since they can produce unrealistic results. Our compilation indicates that the use of a single equation is not adequate to resolve the broad spatial TOC distribution in marine environments and that organic facies modelling software should consider regional aspects in order to produce adequate quantitative estimations of TOC content. Additionally, we used our dataset to investigate the optimal environmental conditions linked to organic-rich sediments in the South Atlantic. Our results show that high productivity, shallow depths, and optimal sedimentation conditions are required for the occurrence of organic-rich sediments. This can be viewed as important factors preconditioning the formation of black shales in the South Atlantic. Nevertheless, we highlight that orbitally-driven changes in hydroclimate and its impact in weathering-derived nutrient input probably played a major role on boosting productivity and preservation of marine organic carbon during sedimentation of high-TOC black shales in the South Atlantic.

Automated summary

The study evaluates TOC content in recent and past sediments of the South Atlantic, finding that available TOC-predictive equations tend to overestimate TOC distribution at continental margins and the edges of the subtropical gyre. Results also indicate better correlation in the eastern South Atlantic compared to the western. Carbon flux equations should be used cautiously in very shallow and high productive systems.