DEM resolution dependencies of terrain attributes across a landscape

Abstract

This paper documents resolution dependencies in terrain analysis and describes how they vary across landform location. Six terrain attributes were evaluated as a function of DEM resolution-slope, plan curvature, profile curvature, northsouth slope orientation, east-west slope orientation, and topographic wetness index. The research highlights the effect of varying spatial resolution through a spatial sampling/resampling scheme while maintaining sets of indexed sample points at various resolutions. Tested sample points therefore coincide exactly between two directly compared resolutions in terms of their location and elevation value. An unsupervised landform classification procedure based on statistical clustering algorithms was employed to define landform classes in a reproducible manner. Correlation and regression analyses identified sensitive and consistent responses for each attribute as resolution was changed, although the tested terrain attributes responded in characteristically different ways. These responses displayed distinguishable patterns among various landform classes, a conclusion that was further verified by a series of two-sample, two-tailed t-tests.

A geomorphology-based approach for digital elevation model fusion – case study in Danang city, Vietnam

“…Aspect is calculated as circular degrees clockwise from 0 to 360 • , and it is therefore difficult to compare quantitatively (Deng et al, 2007). In order to assess the accuracy in aspect as well as slope, unit normal vector (NV) of topographic surface was considered.…”

Accuracy Assessment of Landform Classification Approaches on Different Spatial Scales for the Iranian Loess Plateau

“…One reason lies in the DEM itself as the strongly dissected relief structure is already insufficiently contained in the input dataset. Generally, lower spatial resolutions lead to a smoother representation of the topography and the relief seems in the DEM much gentler and more rounded as it is in reality [31]. Therefore, neither the ASTER GDEM nor the SRTM dataset include enough information to classify the very small landscape structures.…”

“…In addition, it is rarely shown how different scales and input datasets overall work with these approaches. The questions of accuracy and scaling arise in particular for areas with stronger relief variation as such landscapes are more sensitive for changes in digital elevation model (DEM) resolutions [31]. Generally, most experiences indicate that DEM pixel resolutions between 5 m and 30 m are most suitable for landform classifications [32,33].…”

Multiscale Landforms Classification Based on UAV Datasets

“…Scale is predominantly considered a function of the resolution of Digital Surface Models (DSMs) (Hengl & Evans, 2009;Mac Millan & Shary, 2009). The dependency of land surface has been confirmed by several researches (Chang & Tsai, 1991;Wood, 1996;Florinsky & Kuryakova, 2000;Evans, 2003;Hengl, 2006;Arrell et al, 2007;Deng et al, 2007;Pogorelov & Doumit, http://www.scholink.org/ojs/index.php/se Sustainability in Environment Vol. 3, No.…”