Five common mistakes in fluvial morphodynamic modeling

Abstract

Recent years have seen a marked increase in the availability of morphodynamic models and a proliferation of new morphodynamic codes. As a consequence, morphodynamic models are increasingly developed, used and evaluated by non-experts, possibly leading to mistakes. This paper draws attention to five types of common mistakes. First, new morphodynamic codes are developed as extensions of existing hydrodynamic codes without including all essential physical processes. Second, model inputs are specified in a way that imposes morphodynamic patterns beforehand rather than letting them evolve freely. Third, detailed processes are parameterized inadequately for application to larger spatial and temporal scales. Fourth, physical and numerical phenomena are confused when interpreting model results. Fifth, the selection of modelling approaches is driven by 2 the belief that complete data are a prerequisite for modelling and that the application of 2D and 3D m o d e l s r e q u i r e s m o r e d a t a t h a n t h e a p p l i c a t i o n o f 1 D m o d e l s. E x a m p l e s f r o m f l u v i a l morphodynamics are presented to illustrate these mistakes.

A framework for the analysis of noncohesive bank erosion algorithms in morphodynamic modeling

“…The bank erosion coefficient then loses its strict physical meaning and simply becomes a calibration parameter that compensates for the spurious effect of numerical discretization. Failure to recognize and properly interpret such an interplay between physical representation and numerical artefacts is one of the modeling mistakes in morphodynamics categorized by Mosselman and Le []. Excessive mesh‐size dependence also causes practical issues associated with applying the model in grids with varying size cells (as common with triangular or curvilinear meshes in two‐dimensional modeling), and makes it hard to specify suitable default values of the coefficient suitable for general use.…”

Evolution of artificial spawning sites for Atlantic salmon (Salmo salar) and sea trout (Salmo trutta): field studies and numerical modelling in Aurland, Norway

“…6). According to Mosselman & Le (2016), the topography of the riverbed is the most important boundary condition for a morphodynamic model. Furthermore, in the model no corrections according to El Kadi Abderrezzak et al (2016) were carried out with regard to exposure/protection; and therefore, sometimes in the very small test areas (about 50 m 2 ), an increased expression of re-arrangement processes occurs.…”

Numerical simulation of large‐scale bed load particle tracer advection‐dispersion in rivers with free bars

“…According to our trial calculations, a simple interpolation of bed load fluxes computed at other locations (e.g., center of cell or boundary of cell) to x p i and y p i can cause development of very small bars with high transverse mode. This may be because such an interpolation results in the use of many discrete points in computing the local bed slope, leading to inaccuracy in a parameter that plays an important role in the inception of free bars [Kuroki and Kishi, 1984] as well as in the stabilization of the computation of bed evolution [Mosselman and Le, 2016].…”

“…The calculations here are at experimental scale: channel width is 0.48 m, grain size is 1.3 mm, bed slope is 0.075, and water discharge is 3 l/s, corresponding to a Froude number of 0.88, a Shields number of 0.06, and a width-to-depth ratio of 27.7. Mean step length characterizes a Journal of Geophysical Research: Earth Surface 10.1002/2016JF003951 stabilizing effect on bed morphodynamics [Mosselman and Le, 2016]; the longer step length further suppresses the conditions for the linear development of free bars [Kuroki and Kishi, 1984]. Figure A1 shows the sensitivity of wavelength and wave height of free bars to the type of morphodynamic model (flux versus entrainment), and variation in mean step length.…”