WP leader: University of Miskolc, Endre Dobos
We will develop a methodology for a global platform for terrain and soil parent data at a scale of 1:1 million based on existing SOTER criteria. Four activities are planned:
- Development of an artifact-free DEM. DEMs derived from SRTM synthetic aperture radar measurements processed by interferometry suffer from thermal noise. This will be reduced using enhanced filter techniques developed by Selige et al (2006). Reflection of the radar signal by the vegetation canopy also produces a false surface; corrections will be applied using Europe-wide land cover data; more…
- Morphometric characterization of landforms, based on the existing SOTER landform criteria using elevation, slope, relief intensity and dissection parameters, and further developing the digital, SRTM-based procedure of Dobos et al (Dobos et al. 2005); and the creation of landform units;
- Determination of soil parent material within these landform units, based on low-resolution satellite imagery (AVHRR, MODIS, SPOT Vegetation) and DEM data in combination with legacy soil parent material data using a classification of parent materials relevant for soil development and based on the system developed by BGR (2005); and
- Creation of terrain units by combining landform and soil parent material units. Research will focus on the generalization and aggregation processes and on methods of cleaning and structuring the geometric data sets for the four 1:1 million-scale windows.
The end-product of WP 1 will be a 1:1 million-scale SOTER geometric databases of the terrain units for the four windows: combining landform and parent material derived from:
- a quantitative, DEM-based algorithm for landform classification applied on a complete, artifact-free digital elevation model for Europe South of 60 degrees, for Morocco and for South China, and
- methods to derive parent material classes and delineations using RS and legacy data and a simplified soil parent material classification relevant to soil development and performance.