Abstract

Alpine rivers have been regulated to claim productive land in valley bottoms since the last two centuries. Width reduction and rectification often induced the development of regular scour-deposition sequences, called alternate bars, with implications for flood protection, river navigation, environmental integrity. Understanding how alternate bars evolve in rivers and defining the key aspects that influence the development of these regular deposits of sediments represents a challenge that is not fully described. Most studies on alternate bars are in fact based on mathematical theories, laboratory experiments and since 1990s numerical simulations, but only few studies on field cases have been performed so far. The goals of this work are: i) to quantify the morphodynamics of alternate bars in the Alpine Rhine River, with a particular emphasis on bar migration; ii) to assess to what extent the predictions of analytical bar theories are consistent with field observations and to explore how theories may help interpret observed alternate bars dynamics; iii) to determine the ability of a numerical model to simulate correctly the formation and the length scale of alternate bars and the influence of different multi-decadal inflow conditions. The 42 km chosen reach is located along the border between Austria and Switzerland, between the confluences of Landquart and Ill rivers. The whole reach has been completely embanked starting from the 19th century, so alternate bars have been present for more than a century. Moreover the simplification of the cross section, together with the presence of only few bends, puts the Alpine Rhine in the ideal position to be compared with analytical theories of alternate bars in straight channels. The goals are achieved by analyzing a dataset of freely available Landsat imagery, which combine unprecedented temporal length (3 decades), spatial length (more than 400 channel widths) and temporal resolution (around 3 images per year). Bars show a spatially selective behavior, with short bars occurring in distinct straight reaches with respect to longer bars. The same evidence is found in terms of bar migration, so that short bars are shown to migrate more than longer bars, in agreement with theoretical predictions. A full range of bar wavelengths and more complex patterns occur in reaches with bends and ramps. Bar height, obtained from cross section monitoring, was found to be much more uniform. The temporally long dataset, including approximately 30 floods with different magnitude and duration, allowed the investigation of bar migration as a function of discharge, showing that bars migrate faster for intermediate floods. Predicted values of linear theories for free and forced bars in straight channels are in good general agreement with field observations, when considering conditions of bar formation and bar wavelength. Comparing theories and observations suggests that theoretical outcomes may represent the boundaries of the actual, wide range of bar behavior, which likely reflects non-linear interactions, flow unsteadiness, sediment size heterogeneity and finite length of straight reaches, which are not retained in linear theories. Non-linear interactions are investigated through the 2D numerical morphodynamic model Basement, developed at the Swiss Federal Institute of Technology of Zurich. Preliminary investigations focus on the role of the transverse sediment transport, that behaves as a diffusive term. The numerical diffusion can be indirectly evaluated starting from the calibration of the coefficient of the diffusive term. A benchmark methodology to evaluate the lateral and numerical diffusion is defined. The results are used in the morphological calibration of the model. The spatial trend of wavelengths is in general agreement with the field data, and the migration takes place mainly in correspondence to short bars, whereas long bars tend to elongate with time. The choice of a constant discharge or a real hydrograph influences the time scale of bar evolution. The present analysis results in the longest spatial and temporal field case study of river bars in channelized rivers with a temporal survey resolution that allows the investigation of the effect of individual flood events, and provides new quantitative data on bar wavelength and migration. The dataset provides useful information to assess the applicability of analytical bar theories, so far mainly tested against flume experiments, and following recent attempts in French and Dutch streams. Moreover, a novel two-dimensional morphological benchmark to access the role of numerical diffusion is proposed.