Files in this item



application/pdfWATERMAN-DISSERTATION-2017.pdf (41MB)
(no description provided)PDF


Title:Non-cohesive bank migration in meandering rivers and bank accretion in weakly braided rivers
Author(s):Waterman, David Michael
Director of Research:Garcia, Marcelo H.; LaGory, Kirk E
Doctoral Committee Chair(s):Garcia, Marcelo H.
Doctoral Committee Member(s):O'Connor, Ben L; Best, James L.; Parker, Gary; Rhoads, Bruce L.
Department / Program:Civil & Environmental Eng
Discipline:Civil Engineering
Degree Granting Institution:University of Illinois at Urbana-Champaign
Bank erosion
Bank accretion
Abstract:The first topic addressed in the dissertation is the migration of meandering rivers whose banks consist wholly, or partly, of non-cohesive bed-material-sized alluvial deposits; such banks are known as composite banks when the near-surface layers consist of fine-grained deposits. The relevant forces and environmental conditions that must be taken into account to quantify the initial movement and transport of non-cohesive soils are relatively few and more readily quantified as compared to cohesive soils. Thus, the treatment of banks formed of non-cohesive materials would seem to be relatively straight-forward and perhaps resolved long ago. However, existing methods to evaluate channel migration when the banks are formed of non-cohesive material have conspicuous inadequacies. Basic questions have yet to be adequately answered, such as: (a) How fast would a river meander-bend migrate if the migration rate was dictated entirely by the non-cohesive portion of the bank? (b) What is the expected cross-sectional shape of such a channel? Due to the difficulties in establishing freely meandering channels in non-cohesive materials in a laboratory setting, even empirically-based answers to these questions are lacking. The analysis herein attempts to shed light on these issues, under the guiding principle that, until the simplest possible scenario is adequately understood, additional progress regarding the more complex realities of natural streams will be slowed and perhaps misguided. The most straight-forward treatment of non-cohesive channel boundaries is to implement the sediment mass conservation equation (Exner equation) with known sediment transport relations and numerically simulate the temporal evolution of channel deformation. Generalities are difficult to establish from such treatments; and the computational expense of achieving adequate spatial resolution in the near-bank region makes such an approach unfeasible for practical usage. A suitable alternative method is a bank-integrated approach, in which the Exner equation is integrated over the bank region under a shape similarity assumption. However, the proper bank shape to specify is unknown, and the solution for the migration rate is highly dependent on the shape specified. The analysis herein addresses the issues of shape and migration rate using both analytical and numerical techniques, supplemented with past laboratory data and current field observations from the Mackinaw River in Illinois. The analysis indicates that, in order to establish a bank that migrates without changing shape (parallel migration), the bank shape over the region where the critical shear stress is exceeded will be concave upward. Under conditions of uniform, fully-developed bend-flow, the transverse slope at the base of the bank (the channel thalweg) is demonstrated to be zero when considering boundaries consisting of uniform non-cohesive alluvium. The helical flow contribution drives the transverse flux of sediment from the base of the bank, rather than the downslope gravitational force contribution. The bank profile adjusts itself to satisfy both the demand for transverse flux at the base of the bank concurrently with the zero transverse flux boundary condition at the top of the non-cohesive layer to yield parallel migration. The role of curvature in meander bend migration is shown not to be solely the generation of excess boundary shear stress (or excess velocity); of equal importance is its ability to cause transverse flux of bedload away from the base of the bank to allow migration rather than bank slope relaxation. The past idea of unimpeded removal with respect to basal endpoint control is formalized into a rational, dimensionless migration rate formula; it sets an upper bound on migration rate in the absence of any confounding effects such as bank armoring or boundary shear stress modifications (vegetation, bank irregularities) that are generally encountered in real streams. Demonstration is provided that the transverse slope of the channel bed region, typically specified using simple empirical approaches in reduced-order morphodynamics models, is strongly influenced by the channel migration rate. In the presence of a migrating outer bank, the transverse slope may be considerably lower, and the depth at the outer bank considerably less, than the predictions of earlier theoretical models. The second topic addressed in the dissertation is bank accretion in weakly braided streams, with special attention given to bank accretion as a mechanism that can lead to general channel geometry adjustment. Bank accretion in single-threaded meandering streams occurs through the gradual colonization of the high elevation portions of point bars; however, bank accretion in braided streams is a more challenging topic, due to the bars being more dynamic and less predictable. The particular case under consideration in the dissertation is the middle Green River in eastern Utah. Modification of both the hydrology and sediment input to the reach of concern has resulted in channel geometry adjustments since regulation of river flows from the Flaming Gorge Dam began in 1962; however, the presence of a major contributor of bed-material-sized sediment (Yampa River) upstream has limited the geometry adjustment. Potential adjustments that could accommodate the modified sediment feed rate with the given hydrologic regime include longitudinal slope adjustment, reconfiguration of the cross-sectional geometry, modification of the bed grain-size distribution, or a combination. Although the combination of potential adjustments is immense, and not subject to a simple solution, the solution space is analyzed with the intention of providing insight into the general directions of adjustment in the context of the available mechanisms to achieve those adjustments. An understanding of bank accretion requires that the sandbar dynamics be well understood. Bar dynamics are evaluated with multiple objectives. The first objective involves gaining understanding of a particular geomorphic form associated with the sandbars, a depressional feature between the sandbar and river bank that is utilized by Colorado Pikeminnow (Ptychocheilus lucius), an endangered fish species that survives only in the rivers of the Colorado River basin. The second objective involves utilizing knowledge regarding bar dynamics to better understand ongoing bank accretion. Field data obtained in the middle Green River during the 2014 spring flood is utilized, along with sedimentological data obtained later that season, with particular attention given to interpretations regarding sandbar interactions with other sandbars. A conceptual model is presented of three canonical cases of bar interactions in this system that differ based on the degree of channel geometry forcing of the bar positions. The bar interactions cause the sandbars associated with backwater-habitat sites to go through a natural cycle of birth, growth, death, and rebirth; the general lack of long-term stability reduces the likelihood of bars becoming colonized with perennial vegetation and accreting to the bank. In addition to the field investigation, an analysis of 1997 and 2015 aerial photography is presented that illustrates sites of large-scale bank accretion (>20-m maximum width) within the evaluation area. Landsat satellite imagery from 1984 to 2015 is used to identify emergent bar presence during the base flow season, and the bank accretion sites are evaluated with respect to the temporal probability of emergent bar presence. At the evaluation scale of multiple channel widths, a number of statistical metrics are evaluated to describe bar pattern, such as spatial variance and entropy, to identify potential relationships with the bank accretion sites. However, only local statistics at the 30-m by 30-m pixel scale are found to provide a strong relation to the locations of bank accretion. Pixels having a high temporal probability of emergent bar presence during the base flow season are interpreted as depositional areas associated with strong river geometry forcing conditions.
Issue Date:2017-07-10
Rights Information:Copyright 2017 David Waterman
Date Available in IDEALS:2017-09-29
Date Deposited:2017-08

This item appears in the following Collection(s)

Item Statistics