Geomorphometric comparison of DEMs built by different interpolation methods

Abstract

One of the most important stages of creating DEMs is the selection of a suitable interpolation algorithm. In this paper I decided to take a look at the most popular methods of data interpolation: Inverse Distance Weighting, Natural Neighbour, Spline, Radial Basis Functions, Local Polynomial and Kriging. As the research area served fragment (20 km2) of the Silesian Upland with diversified relief. I analysed visual effects (3D view and profiles), summarized the basic geomorphometric statistics (heights, local relief, slopes, aspects, curvatures) and an assessment of the vertical accuracy of developed models (RMSE and result conformity) have made. After conducted studies it can be stated, that the best interpolation methods for analyse of the relief are Natural Neighbour and Kriging, because they do not create any artefacts.