Year of Publication


Degree Name

Doctor of Philosophy (PhD)

Document Type





Civil Engineering

First Advisor

Dr. James Fox

Second Advisor

Dr. Scott Yost


Turbulent structures are known to be important physical processes in gravel-bed rivers. A number of limitations exist that prohibit the advancement and prediction of turbulence structures for optimization of civil infrastructure, biological habitats and sediment transport in gravel-bed rivers. This includes measurement limitations that prohibit characterization of size and strength of turbulent structures in the riverine environment for different case studies as well as traditional numerical modeling limitations that prohibit modeling and prediction of turbulent structure for heterogeneous beds under high Reynolds number flows using the Navier-Stokes equations. While these limitations exist, researchers have developed various theories for the structure of turbulence in boundary layer flows including large eddies in gravel-bed rivers. While these theories have varied in details and applicable conditions, a common hypothesis has been a structural organization in the fluid which links eddies formed at the wall to coherent turbulent structures such as large eddies which may be observed vertically across the entire flow depth in an open channel. Recently physics has also seen the advancement of topological fluid mechanical ideas concerned with the study of vortex structures, braids, links and knots in velocity vector fields. In the present study the structural organization hypothesis is investigated with topological fluid mechanics and experimental results which are used to derive a vortex model for gravel-bed flows. Velocity field measurements in gravel-bed flow conditions in the laboratory were used to characterize temporal and spatial structures which may be attributed to vortex motions and reconnection phenomena. Turbulent velocity time series data were measured with ADV and decomposed using statistical decompositions to measure turbulent length scales. PIV was used to measure spatial velocity vector fields which were decomposed with filtering techniques for flow visualization. Under the specific conditions of a turbulent burst the fluid domain is organized as a braided flow of vortices connected by prime knot patterns of thin-cored flux tubes embedded on an abstract vortex surface itself having topology of a Klein bottle. This model explains observed streamline patterns in the vicinity of a strong turbulent burst in a gravel-bed river as a coherent structure in the turbulent velocity field.