Marine sediments from the deep ocean are determinant in biogeochemical cycles of major elements (C, N, P, Si) on a millenial or greater timescale. The dynamics of biogeochemical reactions in this reservoir are thus of prime importance for the understanding and quantification of the role of the ocean in the transitions during climatic cycles and geological times. The study of surficial sediments involves both field work and modelling. Using two examples, we show how models and data may interact to improve our understanding of surficial sediment dynamics. In the Antarctic Polar Front, a steady state approach provides a picture of the sediment organic matter as a multi-reactivity component and indicates a strong coupling between production and deposition. This is confirmed by the analysis of porewaters (O-2, NO3) and of the solid phase (C-org, Delta(14)C of the organic matter). In the North Atlantic, a non steady state study, using data and models, shows a damped response of the early diagenetic system to seasonal variations in the flux of particulate organic carbon. In contrast, interannual variations, possibly more intense, could prevail over the temporal changes of surficial sediment composition.