Scales and hierarchies in landform classification

Although digital terrain models (DTM) are being extensively used for various modelling purposes, multi-scale hierarchical approaches are still less developed. The overall objective of this project is to provide a coherent framework of nested hierarchical organization of topography for modelling purposes. In order to address this important topic, not only for geomorphometry, but for all fields that need morphometry as a boundary condition, innovative techniques supported by fieldwork will be developed and applied. Framed within multi-scale fuzzy analysis, the methodology will: 1. differentiate characteristic scales associated with specific parameters and/or parameters behaviour across scales; 2. test the significance of morphometric patterns across scales; 3. set up a nested hierarchy of landforms, both morphometric and morphologic relevant. The expected results of this project have a great potential of improving geomorphological and environmental modelling across scales which is an issue that has not been adequately addressed so far. Moreover, the project has the potential of opening up new horizons in robot terrain perception and spatial data mining.

A landform comprises a geomorphological unit, and is largely defined by its surface form and location in the landscape, as part of the terrain. It is an element of the topography of a landscape while the characterisation of landform elements is intrinsically linked to the processes which form them. For instance, alluvial fans or canyons are dominated by fluvial processes. Traditionally, scientists from different disciplines mapped such landforms in the field or based on aerial photographs. This requires domain knowledge and is partially a subjective process. The increasing availability of Digital Elevation Models (DEMs) triggered the search for methods to semi-automatically derive elementary landforms (also: facets or relief units). The SCALA project developed methods to derive homogeneous divisions of the land surface, at the given scale/resolution. These units are generic, i.e. not process specific. They represent relatively homogenous terrain units that are sub-divided by lines of discontinuity in the terrain. Since the spatial distribution of landforms is often scale-dependent, the SCALA project was particularly dedicated to issues of spatial scale and hierarchy. What can be considered as a small hill at one scale might be included in the plateau at a different scale. The hierarchical structure of topography could be demonstrated in several well received journal articles. We showed that models of topography bear hierarchical structures organised around characteristic scales. These characteristic scales can be detected statistically. However, relationships between pattern and processes still remain to be determined. SCALA developed new methods for automated classification of terrain units at global scale and for the estimation of optimal scales for the representation and analysis of such terrain units. The methodology developed could be successfully applied to soil-landscape modelling and to landform delineation and classification for archaeological purposes. SCALA gained worldwide attention when introducing a method to automatically classify optimal scales of terrain units at a global level. A DEM is automatically segmented and classified at three scale levels. For each domain, optimal scales in the DEM are statistically detected and terrain units are automatically created at these scales. The terrain units are sub-divided into smaller terrain units based on thresholds that are defined by the mean values of elevation and variation. Results resemble reasonably patterns of existing global and regional classifications, displaying a level of detail close to manually drawn maps while at the same time allowing for automaton and transferability of the mapping process.