On the development of a digital elevation model over South Africa using ground and satellite data

Malindi, Mihlali

On the development of a digital elevation model over South Africa using ground and satellite data

Master Thesis

2022

Abstract

A digital elevation model (DEM) represents the bare land surface of the Earth. DEMs are used in a wide range of applications, including geological studies, geomorphology, water resources and hydrology, evaluation of natural hazards, and vegetation surveys. In recent years, DEMs have increasingly been used in geographic information systems (GIS), mainly due to the availability of free satellite-based DEMs, some with global coverage. The satellite-based DEMs over South Africa provide topographic surface representation but are associated with errors, and in recent decades there have been significant efforts to improve accuracy. In South Africa, the ground levelling (trigonometrical beacon) data is more capable of representing the terrain heights accurately. However, the data points are farther apart, which makes it difficult for accurate continuous terrain representation. In this research, contributions are made towards the development of an accurate digital elevation model from ground and satellite data over South Africa. This is achieved by preparing satellite-based DEMs (AW3D30, SRTM, ASTER, TanDEM-X, and MERIT), assessing the quality of the satellite-based DEMs, selecting candidate DEMs for fusion, modelling candidate DEM errors, and fusing DEMs. The aerial-based DEM from LiDAR is also applied in the assessment of the quality of satellite-based DEMs, although this was only possible in selected areas due to a lack of LiDAR data covering the whole of South Africa. Following removal of outliers from each DEM, a different number of ground levelling data is used in the assessment of the DEMs (26364, 25728, 23773, 25967 and 24485) ground levelling points for AW3D30, SRTM, ASTER, TanDEM-X and MERIT, respectively. The vertical quality assessment results indicate that the standard deviations of the differences between ground levelling and DEMs heights are ±5.09, ±7.03, ±9.20, ±4.99 and ±8.36 m for AW3D30, SRTM, ASTER, TanDEM-X and MERIT, respectively. In general, the vertical accuracies of the satellite-based DEMs are relatively lower in higher areas than in low areas. The results of height differences between satellite-based and LiDAR DEMs heights in different geomorphological ranges indicate that the AW3D30 and TanDEM-X are better candidate DEMs for generating a new DEM over South Africa. Applying a combination of linear regression, multiple regression, and adaptive terrain-dependent methods to these DEMs, their vertical accuracies improved. The standard deviations of the differences between ground levelling and the improved DEMs at 8,657 points over South Africa decreased from ±5.745 to ±4.995 m for AW3D30 and ±5.073 to ±4.582 m for TanDEM-X. A fused DEM was developed from improved AW3D30 and TanDEM-X DEMs using a combination of different fusion methods (linear combination, weighted averaging, and simple averaging) over South Africa. The fused DEM was assessed using 8,657 ground levelling points over South Africa. The standard deviation of the height differences between ground levelling and the fused DEM is ±4.290 m, indicating the superiority of the fused DEM over all the satellite-based DEMs used in this study. The fused DEM can be applied in areas with a slope less than 20° where an accuracy of less than 4.3 m is achievable. In the steepest areas, it can still achieve better vertical accuracies compared to other satellite-based DEMs tested.

Keywords

Engineering

Reference:

Collections

Masters

Full item page