From source to sink: towards a statistical-numerical model of sediment generation and evolution

Category Other
Group GSI.IR
Location International Geological Congress,oslo 2008
Author von Eynatten, Hilmar; Tolosana-Delgado, Raimon
Holding Date 21 September 2008

Numerical modeling is nowadays a powerful tool in several fields of geosciences. In sedimentology and basin analysis, such modeling usually comprise basin subsidence, heat and fluid flow, basin fill (stratigraphic modeling) as well as climatic, oceanographic and biological forcing of sedimentary systems. But little has been done so far concerning a comprehensive model for describing the composition of clastic sediments, i.e. grain size, mineralogy, geochemistry. The final aim of this research is to obtain a model providing a quantitative description of the petrographic, chemical and granulometric changes occurring at the transfer of material from the original source rock to its final site of deposition in a sedimentary basin. These changes are largely controlled by the source rocks, tectonics, physiography, climate and transport energy.
The first steps of our model focus on two of the most significant processes shaping sediment composition, mechanical comminution and chemical weathering. To model the influence of comminution we present chemical compositions of modern fluvio-glacial sediments that originate from a small and largely homogeneous granitoid catchment area in the Central Alps of Switzerland. Sediment composition is analysed for most of the relevant grain size range from clay to granules in one-phi steps. The dataset is analysed statistically using the Aitchison log-ratio approach, the chemical data are transformed to mineralogical compositions by endmember modelling, and, finally, a log-linear regression model is developed. Results underline that, to a large extent, no chemical processes are needed to generate the observed compositions. Consequently, the differences in the geochemical composition of each grain size fraction are primarily related to comminution and hydrodynamic sorting, i.e. they are controlled by the inherited grain size from the source rocks as well as fissility, hardness, density and shape of individual minerals. This can be best modelled by explaining the composition as a function of grain size with a log-linear trend plus steps at the sand-silt and at the silt-clay thresholds.
If weathering is the most important process, one would expect finer fractions to be more weathered than coarser ones, because: a) weathering is typically considered to be faster in smaller grains due to their higher specific surface; and b) typical weathering products are preferentially enriched in the finer fractions. However, classical feldspar-based weathering indexes are not ideally suited for evaluation of weathering along the full grain-size range, as feldspar is not the dominant component in the fine-grained fractions. Therefore we use a recently published model developed for a wide range of source rocks. Comparisons between the fitted trend and the existing weathering indexes show substantial differences, thus reinforcing our conception of almost absolute lack of weathering in these sediments.