Effect of digital terrain model resolution on topographic parameters calculation and spatial distribution of errors

ABSTRACT: Digital terrain models (DTM) have become a commonly used source of data describing the shape of the Earth. Resolution determines the level of details that can be represented in this model. Changing the resolution influences the accuracy of the DTM. The effect of a resolution is even more clearly seen in slope and aspect calculation. Despite the resolution, there are different methods of acquiring DTM that contribute errors. Each method requires its own algorithms for data management and its own interpolation techniques. Some of the most popular methods are: laser scanning (LiDAR), photogrammetric measurements on pairs of aerial photos or satellite images, digitizing contour lines from topographic maps and InSAR measurements (e.g. SRTM model). Both resolution and acquiring method as well as many other factors constitute the overall error of a DTM. The knowledge of its size is the key information for the DTM user and is required for proper use of a model. Each research requires a specific level of accuracy which depends highly on the nature and scale of the research. For most terrain models errors are estimated with RMSE statistic. It is calculated by comparing cells in DTM with control points representing ‘‘true values’’. The main objective of this research is to investigate how horizontal resolution influences vertical accuracy of a DTM and how it influences calculation of slope and aspect. The general approach is to generate DTM, slope and aspect covering the same area with 1, 5, 10, 25, 50 and 100 m resolution for all four methods of acquiring elevation data. Moreover, the results of analysis of four different models have been compared with LiDAR model. All cell values from this model were used as a reference in statistics calculation. Basic statistics as well as spatial distribution of errors were examined according to different resolutions.