The mineralogical composition of precursor sediments of calcareous rhythmites: a new approach

Previous studies in Silurian carbonates from Gotland (Sweden) have led to a model for the development of limestone–marl alternations. This model postulates that early diagenesis of precursor sediments without strong primary differences can result in a differentiation by selective dissolution of aragonite in marl beds and reprecipitation of calcite cement in limestone beds. This model is described as a set of mathematical equations that quantify the diagenetic processes (aragonite dissolution and calcite reprecipitation) that occur during the formation of limestone–marl interbeds from a hypothetical homogeneous precursor sediment. The calculations demonstrate that resulting hypothetical limestone–marl alternations show characteristic mathematical relationships between the ratios of the bed thicknesses of limestones and marls on one side, and the carbonate contents, on the other. By reversing this model, the original mineralogical composition of the precursor sediment of real-world rhythmic successions can be determined. In this study, alternations from the Silurian of Gotland, the Cambrian, Devonian, and Mississippian of North America, the Jurassic of France and Germany, and the Cretaceous of France are shown to exhibit mathematical relationships similar to those calculated for hypothetical precursor sediments without primary differences. Therefore, the mineralogical composition of their precursor sediments can be estimated. In contrast, the clear mismatch shown by the Lower Jurassic Belemnite Marls from Dorset indicates that these rhythms did not suffer an early diagenetic overprint. Our model helps to differentiate between rhythmites with strong depositional variations and those without; however, it cannot indicate whether a given alternation is the product of rhythmic diagenesis of a homogeneous precursor sediment or the result of diagenetic enhancement of subtle underlying sedimentary rhythms. For horizontally correlated patterns, such as laterally extensive beds and layers of nodules, an a priori unknown external signal has to be assumed