Identification of common error signatures in global digital elevation models based on satellite altimeter reference data

The generation of a near-global set of orthometric height data derived from satellite altimetry has allowed for the first time an independent evaluation of the accuracy of existing Global Digital Elevation Models (GDEMs) on a near-global scale. This paper presents results from an intercomparison of GLOBE v1, GLOBE beta and JGP95E with altimeter based datasets derived from the ERS-1 geodetic mission, together with ERS-1 and ERS-2 35 day data. One result of this research is the identification of common error signatures, where correlated height differences are seen to exist between the altimeter data and several of the DEMs. The ability to identify such erroneous ground truth data using altimetry facilitates the improvement of these global and regional models. Having targetted regions where the height data are inaccurate, it enables examination of these data for mis-registration and reference datum errors, which can then be corrected using altimetry derived heights. (C) 2000 Elsevier Science Ltd. All rights reserved.