Date on Master's Thesis/Doctoral Dissertation

12-2025

Document Type

Doctoral Dissertation

Degree Name

Ph. D.

Department

Civil and Environmental Engineering

Degree Program

Civil Engineering, PhD

Committee Chair

Rockaway, Tom

Committee Member

Ghasemi Fare, Omid

Committee Member

Berson, R. Eric

Committee Member

Mahoney, D. Tyler

Author's Keywords

2D hydrodynamic modeling; stream restoration; civil design; hydraulic modeling

Abstract

In the channels and floodplains of inland waterways the hydrodynamic forces of floods destroy property and infrastructure, threaten lives, and damage aquatic habitat. Flood simulations using robust and efficient depth-averaged two-dimensional hydrodynamic models (2D models) can be interpreted to design effective solutions for areas vulnerable to damage. The reliable interpretation of 2D models for use in design is improved here through direct comparison of 2D model output to observed flood events at 6 headwater research sites. Instrumentation was used for model calibration and to define the surface hydrodynamics at sites with confined flows, wide and shallow flows, at vertical steps, and within tight bends. In tight bends (radius to width ratio of 1.1), 2D model output matched the observed pattern of erosion and diversion of flow into the floodplain. In steps, 2D models effectively predicted the location and intensification of velocity in near-critical flows with maximum (max) Froude (Fr) number of 0.9, hydraulic jumps (max Fr 1.8), and impinging flows (max Fr 2.9). The predicted flow on the floodplain ranged from 0 to 80% at sites and matched the gross distribution visible in flood observation imagery. Reliable interpretation of 2D models involves a shift in the scale of interpretation from individual grid cells in wide and shallow, near-critical flows to short sub-reaches in confined flows, hydraulic jumps with Fr >1.7, impinging flows, and tight bends. The required grid resolution necessary to achieve convergence in prediction of max velocities was dependent on the local geometry and ranged from a ratio of model cells to feature geometry of 0.5 to 5. The eddy viscosity coefficients which maintained high max velocity predictions in the model were between 0.25 to 1.0. Separate criteria for selection of coefficients are recommended for wide and shallow flows and confined flows with an active floodplain. At one site, post-event flood evidence was organized into four categories of flood disturbance, from stable to eroding, with statistically significant differences in model predicted velocities between categories. The category associated with the onset of erosion and damage to vegetation had a median velocity of 1.1 m/s, consistent with published thresholds for instability.

Download

Share

COinS