Braided River Evolution and Bifurcation Dynamics During Floods and Low Flow in the Jamuna River

Research output: Contribution to conferencePaperOther research output

Abstract

River bifurcations have become recognised over the last decade as being critical but poorly understood elements in many channel systems, including braided and anastomosing rivers, fluvial lowland plains and deltas. They control the partitioning of both water and sediment with consequences for the downstream evolution and for river and coastal management. Avulsion studies and bifurcation modelling suggest that symmetrical bifurcations are inherently unstable. However, the simultaneous activity of channels in deltas, anastomosing rivers and large braided rivers such as the Jamuna suggest that symmetrical bifurcations are stable in agreement with sediment transport optimisation theories. These theories are still a matter of debate. Our objective is to understand the stability and evolution of the braid pattern through studying the dynamics of the bifurcations under natural discharge conditions: both during floods and low flow. Using a series of Landsat TM images taken at irregular intervals showing inter-annual variation, we studied the evolution of a large number of bifurcations in the Jamuna river between 1999 and 2004. The images were first classified into water, bare sediment and vegetation. The contiguous water body of the river was then selected and translated into a network description with bifurcations and confluences at the nodes and interconnecting channels. Channel width is a crucial attribute of the network channels as this allows the calculation of bifurcation asymmetry. The key step here is to describe river network evolution by identifying the same node in multiple subsequent images as well as new and abandoned nodes, in order to distinguish migration of bifurcations from avulsion processes. Nodes in two subsequent images were linked through distance and angle of the downstream connected channels. Once identified through time, the changes in node position and the changes in the connected channels can be quantified Along the entire river the well-known total and active braiding indices were derived from the network. The development of bifurcation asymmetry is assessed by comparing the width asymmetry of each bifurcation in two different, linked scenes. The formation of a dominant channel is observed during persistent, constant low flow condition. A vast majority of the bifurcations in the dataset showed an increase in bifurcation asymmetry during persistent flow conditions, causing a decrease of the number of parallel channels and the formation of one or only a few dominant channels. This is in agreement with bifurcation instability found in bifurcation studies (de Haas and Kleinhans, this conference). The braided character of the river is maintained due to the reactivation of abandoned channels and a reset of bifurcation asymmetry during high discharges in the rain season, causing an increase in the number of channels. We conclude that, in order to counteract the inherent instability of bifurcations, the prevalence of parallel channels in a braided river is the result of bifurcation re-opening forced by discharge peaks.
Original languageEnglish
PagesEP31C-0749
Publication statusPublished - 1 Dec 2010
EventAmerican Geophysical Union Fall Meeting 2010 - San Francisco, USA
Duration: 1 Jan 2010 → …

Conference

ConferenceAmerican Geophysical Union Fall Meeting 2010
CitySan Francisco, USA
Period1/01/10 → …

Bibliographical note

American Geophysical Union 2010 Fall meeting.[EP] Earth and Planetary Surface Processes, [EP31C] The Morphodynamics of Big Rivers: What Do and Don't We Know? II Posters. EP31C-0749

Keywords

  • [1821] HYDROLOGY / Floods
  • [1825] HYDROLOGY / Geomorphology: fluvial
  • [1855] HYDROLOGY / Remote sensing
  • [1856] HYDROLOGY / River channels

Fingerprint

Dive into the research topics of 'Braided River Evolution and Bifurcation Dynamics During Floods and Low Flow in the Jamuna River'. Together they form a unique fingerprint.

Cite this