A database for the digitization of the sedimentary architecture of fluvial systems: uses in pure and applied research

A relational database has been devised as a tool for the digitization of features relating to the sedimentary and geomorphic architecture of modern rivers and ancient fluvial successions, as derived from either original field studies or published examples. The system has been designed in a way that permits the inclusion of hard and soft data – comprising geometries and spatial and hierarchical relationships – referring to classified genetic units belonging to 3 different hierarchical levels, and assigned to stratigraphic volumes that are categorized in terms of deposystem boundary conditions and descriptive parameters. Several applications of the quantitative information generated through database interrogation have been explored, with the scope to demonstrate how a database methodology for the storage of sedimentary architecture data can be of use for both pure and applied sedimentary research. Firstly, an account is given of how the system can been employed for the creation of quantitative fluvial facies models, which summarize information on architectural styles associated with classes of depositional systems. The value of the approach is shown by contrasting results with traditional qualitative models. Secondly, database output on large-scale fluvial architecture has been used in the context of a comparative study aiming to investigate the role of basin-wide aggradation rates as predictors of fluvial architectural styles. The results contrast with what might be expected by commonly considered stratigraphic models; the main implication is the necessity to reconsider continental sequence stratigraphy models or their domain of applicability. This application further provides an example of how the methodology could be generalized to the study of the sensitivity of architecture to its controls. Thirdly, database output has been used to conduct a re-evaluation of previously-proposed approaches to the guidance of well-to-well correlations of subsurface fluvial channel bodies, applied in earlier studies. Making use of the same analogue information, a new probabilistic approach has been proposed as a way to inform or rank correlation panels of channel bodies across equally-spaced wells. Finally, the value of the system as an instrument for constraining object- and pixel-based stochastic structure-imitating models of fluvial sedimentary architecture is collectively demonstrated through a range of example applications employing database output.