< Back to previous page

Project

A cost-efficient methodology for quantitative provenance analysis and mass budget studies in clastic sedimentary geology (southern North Sea Basin).

Sedimentary provenance analysis attempts to uncover the different sources of a sediment deposited in a certain sedimentary basin, recently or in geological history. Different aspects of the sediment composition, such as bulk petrography, heavy mineral composition or single grain geochemistry can provide clues to determine the source areas. Based on these data, a quantitative mixing model can be constructed describing the proportion of sediment derived from each source area. If information on the geometry of the basin and the sediment routing system is available, a sediment budget can be constructed describing the absolute volume of sediments transported into the basin from each source area. If such an analysis is done for multiple time slices during a certain time period, interpretations can be made on the temporal variations in sediment provenance and paleogeography.

This study sets four main goals: (1) A user-friendly and efficient statistical methodology is created to handle large compositional datasets to handle questions on sediment provenance. (2) A new approach for the partitioning of compositional data into a signal related to hydraulic sorting and a signal related to provenance is proposed. (3) A comprehensive quantitative provenance model is constructed based on a multi-proxy approach and the inclusion of a 3D subsurface model to achieve absolute sediment volumes. (4) These approaches are applied to Miocene, and some Pliocene, sediments of the southern North Sea Basin in order to gain more insight into the shifts in provenance and paleogeography in this basin throughout the Miocene.

Compositional data cannot be analyzed with classic statistical methods because of their constant-sum property and non-negativity, yet this is often overlooked. In this study, the data are first log-ratio-transformed so they no longer suffer from the same problems and can be analyzed correctly with classic statistical methods. Cluster analysis is applied to get a first idea of the groups present in the data. Principal component analysis (PCA) reveals the relationship between the different components of the dataset, for example the different heavy minerals, and provides information on which components contribute most to the formation of the observed groups. Based on the PCA, log-ratios are constructed which describe a maximal amount of variation. The log-ratio plots are the basis for provenance interpretations. Compared to the classic approach using cumulative diagrams it is much easier and faster to recognize groups in an objective manner and more data and variation can be captured on a single plot. Finally, a single provenance proxy extracted from the data can be used for spatial interpolation, improving insight into the spatial and temporal variation in provenance.

Changes in sediment provenance only account for a certain portion of the observed compositional variation in any sedimentary deposit. Other common influences on sediment composition are (chemical) weathering, and hydraulic sorting. The influence of weathering is difficult to quantify, but hydraulic sorting processes can be linked to grain size distribution. In this study, a partial-least-squares regression model is constructed of the heavy mineral composition as a function of grain size distribution. Such a model allows for the recognition of the relation between the components of both datasets. In this case, tourmaline, staurolite and the aluminosilicate polymorphs are correlated with coarser grain size classes, whereas epidote, amphibole and garnet are correlated with finer grain size classes.  Based on the resulting model, a predicted heavy mineral composition as a function of grain size distribution and a residual heavy mineral composition, independent of grain size, are calculated. For the Miocene sediments of the southern North Sea Basin, the variation in the residual composition is similar to the variations observed in the total dataset, confirming that provenance interpretations made are not primarily related to hydraulic sorting. Part of the variation in grain size itself can be linked to provenance as the northern marine sediments are finer grained than the southern continental sediments. It should be considered in other studies as well that a correlation between composition and grain size is not automatically the result of hydraulic sorting.

Different provenance proxies may record different aspects of the provenance signal. Very high-resolution proxies, such as single grain geochemistry, may provide detailed information but they may also miss certain source areas if those areas have a very low fertility of the mineral under study.  Due to such issues it is crucial to base a provenance model on multiple proxies. Heavy mineral analysis and grain size analysis are combined with geochemistry of the bulk sediment, fine fraction and glauconite fraction, and zircon U/Pb geochronology in this study, all coupled with dinoflagellate cyst biozonation to constrain the chronostratigraphical age of the studied sediments.  Based on the heavy mineral composition, there is an increase in aluminosilicate polymorphs, staurolite and tourmaline relative to epidote, garnet and amphibole from the northwest to the southeast
of the study area and from the early Miocene to the Pliocene. This trend is related to an increased input of southern continental sediments, also recognized in previous studies. Based on zircon U/Pb geochronology, however, the age distributions in the Campine Basin are very similar throughout the Miocene which can be explained by intensive recycling of sediments, not easily recognized based on heavy mineral composition. The Meuse, draining the Ardennes, provided a large amount of weathered Paleogene sediment cover and eroded Paleozoic sedimentary rocks to the study area. An increased MnO content in the bulk and fine fraction geochemistry can also be related to the influx of weathered sediments from the Ardennes. Recycling within the Campine Basin from the early Miocene to Tortonian can be recognized as well based on the reworked glauconite geochemistry. Only in the late Miocene and Pliocene did the Rhine become the most important contributor of sediment in the Ruhr Valley Graben, with a clear southern zircon age signature. The northern marine sediment input can be interpreted as a background sedimentation as it remained at a rather constant low rate. The input of southern sediments, on the other hand, strongly increased from the early Miocene to the late Miocene and Pliocene.

This study demonstrates the necessity of a multi-proxy approach to construct a comprehensive quantitative provenance model and increases the knowledge on sediment provenance in the Miocene of the southern North Sea Basin, placing larger focus on sediment recycling and import of weathered sediments from the Ardennes.

Date:1 Oct 2015 →  2 Oct 2019
Keywords:mass budget studies, clastic sedimentary geology, southern North Sea Basin, quantitative provenance analysis, cost-efficient methodology
Disciplines:Geology
Project type:PhD project