Methods for Analysing Bedform Geometry and Dynamics

Abstract

Rhythmic bedforms of different spatial scales are common in the marine environment. In shallow seas, the monitoring of bedforms is important because changes in morphology may interfere with offshore infrastructure and navigation. In addition, investigating bedform dynamics improves our understanding of the processes that cause their behaviour. Quantified bedform characterization also contributes to the validation of morphodynamic models. Modern, high-precision and high-resolution bathymetric data enable the detailed analysis of bedforms. Several semi-automated methods (e.g. geostatistical and spectral techniques) have been developed to quantify the geometry (size and shape) and dynamics (growth, change in shape and migration) of subaqueous bedforms. An overview of these different approaches is given and differences in use and potential are described.

An automated procedure to calculate the morphological parameters of superimposed rhythmic bedforms

“…That can, in principle, be achieved by manual measurements along 1D profiles or surfaces, but this is time‐consuming and difficult when dealing with intricate bedforms. For this reason several of the existing semi‐automated methods have not been fully automated (Van Dijk and Lindenbergh, 2017).…”

“…First, a modified DFT method based on Cazenave et al (2013) is used because it is more intelligent and applicable to complex bedforms. Second, an algorithm is proposed to automatically determine wavelengths of interest because the performance of the DFT method is dependent on the selection of a wavelength as input parameter (Van Dijk and Lindenbergh, 2017).…”

“…Wavelet transform is another spectral analysis method for quasi‐harmonic forms that has in the past been applied to characterize bedform morphology (Cataño‐Lopera et al , 2009; Fraccascia et al , 2011; Singh et al , 2011; Kalbermatten et al , 2012; Gutierrez et al , 2013). Its advantage is that it is less prone to the effects of sample size and that it can handle more irregular signals (Van Dijk and Lindenbergh, 2017). It identifies the spectral characteristics of bedforms and, in combination with a series of robust spline filters applied to the data profile, was able to discriminate three bedform scales (Gutierrez et al , 2013).…”

“…Determination of the wavelength of interest An automated algorithm is proposed to determine the wavelength of interest (λ). Most methods, including the Kriging method (Van Dijk et al, 2008), the DFT method (Cazenave et al, 2013), and the wavelet method (Gutierrez et al, 2013), are dependent on manually chosen wavelengths as input parameters (Van Dijk and Lindenbergh, 2017). In the present study an algorithm based on the modified DFT method is proposed to determine (1) bedform superimposition and (2) the scales of bedform wavelengths.…”

“…In addition, two‐dimensional discrete Fourier transform (2D DFT) combined with Butterworth high‐pass filtering can effectively eliminate the regional topographic overprint (Cazenave et al , 2013). However, the spectral leakage of this method tends to underestimate dune height (Van Dijk and Lindenbergh, 2017). Although wavelet transform is generally applied to 1D profiles, it also performs very well in separating dunes of different scales (Gutierrez et al , 2013).…”