| dc.description.abstract | Meandering river floodplains exhibit periodic structures which can be seen in features such as meander bends, point bars, and oxbow lakes. To improve our understanding and better analyze floodplain landscapes created by the dynamics of meandering rivers, characteristic scales need to be identified. Although methods that involve manual measurements of certain floodplain features are of utility, they are limited in their application and are typically very time intensive. Spectral analysis techniques represent an improved approach. For this research, two separate 2D spectral analysis techniques were used: the Fourier transform and the continuous wavelet transform. By using an appropriate theoretical red-noise background spectrum for the landscape, the spectral analysis techniques could provide a power spectrum which is then used to clearly identify the global and local characteristic scales. The results from the analysis of synthetic test images demonstrated such capability of both methodologies, and indicated that both performed similarly although the wavelet transform provides spatial information in addition to scale. The methodologies were then applied to simulated meandering river floodplain of meandering river where two ranges of characteristic scales were identified that corresponded to bend-scale and meander-train scale features. The characteristic meander-scale features also correlated with the surface metrics focal mean, the average elevation within a given area, and rugosity, the ratio between surface area of a given area and the surface area of a completely flat surface. The results show that the spectral analysis techniques can identify characteristic scale of a meander-river floodplain and that the relationship to the surface metrics indicate that it provides information to the topographic structure of the floodplain. | en |
