| An experimental investigation examines the controls on the geometry of cross-sets formed by subaqueous dunes (aggradation rate, distribution of dune height and length, and temporal changes in the geometry and migration rate of individual dunes).; 34 experiments were conducted in two flumes of different size, spanning most of the field of dune existence. Data consist of time series of dune height and length, migration rate, trough-scour depth, and of cross-set thickness for a range of depositional conditions. Cross-set thickness and length were also measured on sediment peels. Data were used to explore and develop recent theoretical models for predicting dune geometry from cross-sets characteristics, and vice versa. Quantitative analysis of temporal changes in geometry and kinematics of individual dunes led to development of a 2-D stochastic model of dune migration and formation of cross-sets.; The effect of aggradation rate on cross-set geometry is usually negligible. The ratio of variability of dune height/variability of dune-trough scour below mean bed level is critical to models. Mean dune height commonly approximates 2.9 x mean cross-set thickness. Experimental data relating cross-set thickness and dune height agree well with data from natural rivers. Because dune geometry is related to flow depth, these models are potentially important for paleoenvironmental reconstruction.; Individual migrating dunes commonly increase in height by trough scouring, and occasionally by overtaking the downstream dune. Typically, dunes eventually decrease in height and disappear as they finally come under the influence of an upstream dune. All types of change in dune geometry affect the 2-D geometry of cross-sets. However, 2-D geometry of cross-sets does not vary systematically with flow velocity. Rates of change in dune geometry and migration rate vary in such a way with flow velocity that dune trough trajectories remain the same.; Computer 2-D stochastic simulated cross-set shape and thickness are comparable to those observed in sediment peels, indicating a good empirical understanding of the variability of dune-trough trajectories. However, in this, mean dune height and cross-set length are larger than expected, because dune interactions leading to dune destruction were not considered. This issue will require further research.*; *This dissertation includes a CD that is compound (contains both a paper copy and a CD as part of the dissertation). The CD requires the following applications: Windows 95/98, Fortran.90, Media Player. |
