The sensitivity of a volcanic flow model to digital elevation model accuracy: experiments with digitised map contours and interferometric SAR at Ruapehu and Taranaki volcanoes, New Zealand

• Published in: Journal of volcanology and geothermal research 2003, Vol.119 (1), p.89-105
• Main Authors: Stevens, N.F, Manville, V, Heron, D.W
• Format: Article
• Language: English
• Subjects: Applied geophysics; Crystalline rocks; debris avalanche; digital elevation model; Earth sciences; Earth, ocean, space; Exact sciences and technology; Igneous and metamorphic rocks petrology, volcanic processes, magmas; interferometric synthetic aperture radar; Internal geophysics; lahar; LAHARZ flow model; Ruapehu; Solid-earth geophysics, tectonophysics, gravimetry; Taranaki/Mt Egmont; topographic map
• ISSN: 0377-0273; 1872-6097
• DOI: 10.1016/S0377-0273(02)00307-4

A growing trend in the field of volcanic hazard assessment is the use of computer models of a variety of flows to predict potential areas of devastation. These can be compared against historic and geologic evidence of past events for model calibration, or used to construct hazard zone maps for mitigation and planning purposes. The accuracy of these computer models depends on two factors, the nature and veracity of the flow model itself, and the accuracy of the topographic data set over which it is run. All digital elevation models (DEMs) contain innate errors. The nature of these depends on the accuracy of the original measurements of the terrain, and on the method used to build the DEM. In this paper we investigate the effect that these errors have on the performance of a volcanic flow model designed to delineate areas at risk from lahar inundation. The model was run over two DEMs of southern Ruapehu volcano derived from (1) digitised 1:50 000 topographic maps, and (2) airborne C-band synthetic aperture radar (SAR) interferometry obtained using the NASA AIRSAR system. On steep slopes of ∼4° or more, drainage channels are more likely to be incised deeply, and flow paths predicted by the model are generally in agreement for both DEMs despite the differing nature of the source data. Over shallow slopes (∼