Impacts of geoid height on large-scale crustal deformation mapping with InSAR observations

An ellipsoidal height datum (e.g. WGS84) is required in SAR Interferometric processing, whilst DEMs are often referenced to a geoid datum (e.g. EMG96). Therefore, geoid height error is introduced due to the inconsistency of height datums in InSAR derived displacement maps. In this paper, the relationship between geoid heights and uncertainties in InSAR displacements is investigated using 27 scenes of Envisat ASAR ScanSAR images and 44 scenes of Image-mode images covering the western Altyn Tagh Fault, northern Tibetan Plateau. A typical error of 8.8 m in SRTM DEM leads to 3 mm uncertainty in an InSAR interferogram with a perpendicular baseline of 100 m, whilst a geoid height of 20 m can result in 10 mm uncertainty. Geoid height varies from one place to another, and is highly correlated with topography. The impacts of geoid height on InSAR displacements increase with perpendicular baselines, but SAR images with a small perpendicular baseline may not be available in some cases. In addition, it may also not always be feasible to use a best-fitting plane to remove the impacts of geoid height. It is thus highly recommended in this paper that DEMs with an ellipsoidal height datum should be employed in interferometric processing, particularly when long-wavelength crustal deformation is targeted.