This paper presents a robust new approach using artificial neural networks in the form of a Self Organizing Map (SOM) as a semi-automatic method for analysis and identification of morphometric features in two completely different environments, the Man and Biosphere Reserve “Eastern Carpathians” (Central Europe) in a complex mountainous humid area and Yardangs in Lut Desert, Iran, a hyper arid region characterized by homogeneous repetition of wind-eroded landforms.
The NASA Shuttle Radar Topography Mission (SRTM) has provided Digital Elevation Models (DEM) for over 80% of the land surface. Version 3.0 SRTM data provided by the CGIAR-CSI GeoPortal are the result of substantial editing effort on the SRTM DEM produced by NASA. Easy availability of SRTM 3 arc second data promoted great advances in morphometric studies and numerical description of terrain surface features as shown by many literature references. The goal of this study was to develop a new semi-automatic DEM-based method for geomorphometric feature recognition and to explore the potential and limitation of SRTM 90 meter data in such studies.
The 3 arc seconds data were re-projected to a 90 m UTM grid. Bivariate quadratic surfaces with moving window size of 5×5 were fitted to this DEM. The first derivative, slope steepness and the second derivatives minimum curvature, maximum curvature and cross-sectional curvature were calculated as geomorphometric parameters and were used as input to the SOMs. Different learning parameter setting, e.g. initial radius, final radius, number of iterations, and the effect of the random initial weights on average quantization error were investigated. A SOM with a low average quantization error was used for further analysis. Feature space analysis, morphometric signatures, three-dimensional inspection and auxiliary data facilitated the assignment of semantic meaning to the output classes in terms of geomorphometric features. Results are provided in a geographic information system as thematic maps of landform entities based on form and slope.
Geomorphometric features are scale-dependent and the resolution of the DEM limits the information, which can be derived. Mega-Yardangs, an aeolian landform due to intensive wind erosion, cover a large area in the hyper-arid Lut desert. They form elongated depressions and ridges with a width ranging from a few meters to hundreds of meters. These features can clearly be recognized and classified when their width is significantly larger than the DEM resolution but become unrecognizable if their width is less than the grid resolution.
Eastern Carpathians are characterized by elongated ridges and valleys. These ridges are dissected resulting in small patches of landforms like ridge with very steep slope, channels with moderate slopes and plains with gentle slope. These local landform patches are large enough to be recognized in morphometric parameters calculated with a 5×5 window, corresponding to 450 m on the ground. With much larger window sizes, these local features disappear and replaced by features describing regional
patterns and ridges. Therefore, size, shape and pattern of identified geomorphometric objects depend on the underlying DEM resolution and the selected window size.
The results demonstrate that a SOM is an efficient scalable tool for analyzing geo-morphometric features as meaningful landforms under diverse environmental conditions. This method provides additional information for geomorphologic and landscape analysis even in inaccessible regions and uses the full potential of morphometric characteristics.