High spatial resolution topographic data are crucial as a prerequisite for evaluating and modelling volcanic hazards. Remote sensing now provides some of the best methods of retrieving digital elevation models (DEMs) over extensive volcanic regions. Here we compare the advantages and limitations of Shuttle Radar Topography Mission (SRTM) DEMs, derived from radar interferometry, and Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) DEMs, derived from digital photogrammetry. Applications are presented for Mauna Kea (MK), Hawaii and Oldoinyo Lengai (OL), Tanzania. We quantitatively assess the accuracy of the respective DEMs to document the size of moderate-sized volcanic features. ASTER DEM accuracy depends on spectral contrast within the image and on the availability of high-quality ground control points. Unlike the ASTER DEMs, which are user-derived, processed SRTM DEMs are provided without estimations of the vertical accuracy for each scene, and the end-user has no control over the processing method. From comparison with a 10-m spatial resolution DEM derived from 1 : 24 000 scale topographic maps of MK, we estimate root mean square errors at 8, 10 and 13 m for SRTM 30-m, SRTM 90-m and ASTER 30-m DEMs, respectively. For scoria cones (<200 m high, <2 km basal diameter), SRTM 30-m, SRTM 90-m and ASTER 30-m DEMs underestimate cone height by 9.5, 27 and 14%, respectively, mostly because of the averaging effect of decreasing spatial resolution. For height estimations of volcanic features higher than ∼100 m in the OL region, all of the DEMs tested were found to be consistent.
Mercator and Ortelius Research Centre for Eruption Dynamics, Department of Geology and Soil Sciences, Ghent University, B-9000 Ghent, Belgium 2:
Remote Sensing and Photogrammetry Unit, Department of Geography, Ghent University, B-9000 Ghent, Belgium