A new quality validation of global digital elevation models freely available in China. Issue 351 (1st November 2016)

• Record Type: Journal Article
• Authors: Li, P.; Li, Z.; Muller, J.-P.; Shi, C.; Liu, J.
• Is Part Of: Survey review. Volume 48:Issue 351(2016)
• Page Start: 409
• Page End: 420
• Publication Date: 2016-11-01
• Subjects: Global DEM -- CGIAR-CSI SRTM v4.1 -- NASA SRTM v3 -- ASTER GDEM v2 -- SRTM X-SAR DEM -- DEM Interoperability; Surveying -- Periodicals; Great Britain -- Surveys -- Periodicals; 526.9
• DOI: 10.1179/1752270615Y.0000000039
• Languages: English
• ISSNs: 0039-6265

Abstract

Abstract : Global Digital Elevation Models (DEMs) are widely used in the study of natural hazards and environmental change on a global scale. This paper focuses on validation of the most widely used and newly released global DEMs (SRTM v4.1, NASA SRTM v3, SRTMX and ASTER GDEM v2) in China. Authors use independent and precise ground GPS observations to assess their absolute accuracies. The SRTMX DEM performs best with a height RMSE of 9.7 m, while the RMSE of the ASTER GDEM2 is slightly better than that of the SRTM v4.1, and SRTM v3 in these regions have the largest RMSE of 16.6 m. However, systematic negative bias still exists in all the global DEMs. Results of the raster-based comparisons between the DEMs are dependent on the knowledge of vegetation type, density and structure to a large extent as well as accurate co-registration. Slope comparisons exhibit a hierarchical slope difference of about 2° between the SRTMX DEM, ASTER GDEM2 and SRTM C-band DEM (v4.1 and v3). This paper provides the first direct evidence and measurement of the product quality of SRTMX and SRTM v3 DEM in China and also offers a benchmark for the future evaluation of following global DEM products.