Determination of the Impacts of Landscape Offsets on the 30-metre SRTM DEM through a comparative analysis with Bare-Earth Elevations
Peter Nwilo, Emmanuel Ayodele & Chukwuma Okolie
Digital elevation models (DEMs) are fundamental spatial data infrastructure that support a wide range of applications in environmental modelling. However, in landscapes covered by vegetation and man-made structures, the DEMs record systematically too-high heights. This is because these obstructing features on the landscape tend to block a portion of the satellite pulses from reaching the ground, thus introducing gaps in the data. This effect also masks the true performance of the DEM in measuring the elevations in open-terrain conditions adjacent to such areas. On DEMs, these features manifest as abrupt transitions in-between open-terrains. The presence of these offsets limits the use of DEMs for many analytical operations where height of the bare-earth is a requirement. Extracting the terrain height component from the data in areas influenced by such cover is a challenging task. This paper investigates the effects of these vertical offsets on the 30-metre DEM from SRTM (Shuttle Radar Topography Mission) in three study sites located in Lagos and Ogun States of South-West Nigeria. The offsets were determined by comparing the SRTM heights with a reference DEM interpolated from local topographic maps of the area. The results showed varying misrepresentations in SRTM elevations with significant differences in the different landscape categories such as bare lands and built-up areas. However, a strong positive correlation was observed between the SRTM DEM and reference DEM with the highest correlation in bare lands (R2 - 0.98) and lowest in wetland forests (R2 ? 0.85). The study also showed that a significant portion of absolute offsets (47 - 70%) in all landscape categories were in the 1-5�� range. This information serves as a valuable resource in efforts directed at developing a bare-earth DEM from satellite-derived elevation datasets, and thus a justification for further research on the impacts of landscape offsets on the SRTM DEM to improve the accuracy and reliability of the data.
Event: FIG Working Week 2017 : Surveying the World of Tomorrow : From Digitalisation to Augmented Reality
Only personal, non-commercial use of this document is allowed.