Abstract
Rivers are corridors of human civilizations. They play a crucial role in sustaining ecosystems, providing freshwater for agriculture and human consumption, and facilitating transportation and trade. They are also dynamic features on Earth: the channels of many rivers often migrate laterally over time. To understand how river channels migrate, many numerical models have been established. Yet there is a lack of methods to systematically test these models. In addition, existing theories are more developed for single-threaded, meandering rivers; for multi-threaded, braided rivers, there are no comparable methods to describe and predict their changes over time. This dissertation aims to bridge these knowledge gaps using the combination of data analyses with numerical modeling. Focusing on evaluating the success of existing theory for meandering rivers, Chapter 2 introduces a new framework that integrates statistical, spatial, and temporal perspectives for testing numerical models of meandering channel migration, and uses the framework to compare data of channel migration of 110 natural rivers with the result of a widely-used channel migration model. Chapter 3 uses numerical experiments to establish a new geometric constraint on the characteristic timescale of the morphodynamic feedback between channel geometry and migration for meandering rivers. To fill observational gaps in addressing changes in braided river channels, Chapter 4 develops an approach that can track the motion of individual channel threads from subsequent remotely sensed images, connecting measured rate of lateral migration with characteristics of the channel threads. Chapter 5 summarizes the main findings, implications, and future directions of this dissertation. Collectively, this work points to new pathways for improving the understanding of how river channels migrate, leading to better assessments of erosion and flooding hazards in the present, interpretation of sedimentary records in the past, and predictions of landscape evolution in the future.
