Research on meandering rivers has a long tradition. For decades, one of the most popular topic among river morphodynamics, has been related to the planform dynamics of different types of meandering rivers, with modeling approaches developing at increasing pace (Seminara, 2006). Recent modelling studies on freely evolving meandering rivers have rather well explained many features of how meanders evolve, explicitly connecting forms and their evolution to processes. However, some aspect concerning specific meander types (e.g. Nicoll and Hickin, 2011), or on specific processes controlling the planform evolution of meandering rivers (Zolezzi et al., 2008) are still incomplete, with gaps that are still relevant especially in linking modelling approaches with observations from real rivers. This PhD thesis addresses some of these open gaps. The first part of this thesis (Chapter 3) is dedicated to a modelling investigation of confined meandering rivers, i.e., meandering rivers that are laterally confined in their planform evolution by geological constraints. For this particular type of meandering rivers, very few observational and modelling studies have been presented so far. The focus of the second part is the testing of some Machine Learning techniques, whose application is quite novel in morphodynamics, for the prediction and analysis of the planform dynamics of meandering rivers. In this second part, classical, analytical meander modelling approaches are integrated with machine learning methods, whose application is new to the field, and with the outputs of remote sensing analysis on the multitemporal dynamics of real meandering river systems. Chapter 4 focuses on a preliminary analysis of the feasibility of predicting the planform evolution of an idealized (modelled) meandering channel using different Machine Learning techniques, while Chapter 5 presents an application of Machine Learning techniques to a set of real meandering rivers to investigate the possible occurrence of the phenomenon of 2D morphodynamic influence in real rivers. The originality of such investigations mainly relies on the fact that 2D morphodynamic influence has been confirmed from theoretical and laboratory investigations but not from the behavior of real rivers. This introductory section presents general concepts about the main subject of the work, about meandering rivers, especially in relation to their morphological and morphodynamic characteristics. It frames the meandering river pattern in the broader discourse on river channel patterns, then describes the different types of meanders, and recalls some key modelling methods that have been used so far to predict their evolutionary dynamics. Before Chapters 3, 4 and 5 that illustrate the research outputs of the thesis, the introduction is followed by a general methods chapter reporting the key elements of the main methodological approaches that have been used, from analytical meander models to machine learning methods.
Using machine learning and analytical modelling to predict the planform dynamics of meandering rivers / Amini, Hossein. - (2022 Nov 07), pp. 1-119. [10.15168/11572_356681]
Using machine learning and analytical modelling to predict the planform dynamics of meandering rivers
Amini, Hossein
2022-11-07
Abstract
Research on meandering rivers has a long tradition. For decades, one of the most popular topic among river morphodynamics, has been related to the planform dynamics of different types of meandering rivers, with modeling approaches developing at increasing pace (Seminara, 2006). Recent modelling studies on freely evolving meandering rivers have rather well explained many features of how meanders evolve, explicitly connecting forms and their evolution to processes. However, some aspect concerning specific meander types (e.g. Nicoll and Hickin, 2011), or on specific processes controlling the planform evolution of meandering rivers (Zolezzi et al., 2008) are still incomplete, with gaps that are still relevant especially in linking modelling approaches with observations from real rivers. This PhD thesis addresses some of these open gaps. The first part of this thesis (Chapter 3) is dedicated to a modelling investigation of confined meandering rivers, i.e., meandering rivers that are laterally confined in their planform evolution by geological constraints. For this particular type of meandering rivers, very few observational and modelling studies have been presented so far. The focus of the second part is the testing of some Machine Learning techniques, whose application is quite novel in morphodynamics, for the prediction and analysis of the planform dynamics of meandering rivers. In this second part, classical, analytical meander modelling approaches are integrated with machine learning methods, whose application is new to the field, and with the outputs of remote sensing analysis on the multitemporal dynamics of real meandering river systems. Chapter 4 focuses on a preliminary analysis of the feasibility of predicting the planform evolution of an idealized (modelled) meandering channel using different Machine Learning techniques, while Chapter 5 presents an application of Machine Learning techniques to a set of real meandering rivers to investigate the possible occurrence of the phenomenon of 2D morphodynamic influence in real rivers. The originality of such investigations mainly relies on the fact that 2D morphodynamic influence has been confirmed from theoretical and laboratory investigations but not from the behavior of real rivers. This introductory section presents general concepts about the main subject of the work, about meandering rivers, especially in relation to their morphological and morphodynamic characteristics. It frames the meandering river pattern in the broader discourse on river channel patterns, then describes the different types of meanders, and recalls some key modelling methods that have been used so far to predict their evolutionary dynamics. Before Chapters 3, 4 and 5 that illustrate the research outputs of the thesis, the introduction is followed by a general methods chapter reporting the key elements of the main methodological approaches that have been used, from analytical meander models to machine learning methods.File | Dimensione | Formato | |
---|---|---|---|
Final+version (1).pdf
embargo fino al 07/11/2024
Tipologia:
Tesi di dottorato (Doctoral Thesis)
Licenza:
Tutti i diritti riservati (All rights reserved)
Dimensione
17.33 MB
Formato
Adobe PDF
|
17.33 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione