The influence of scene selection on height change rates from TanDEM-X DEM differencing - results of the RAGMAC intercomparison exercise

Abstract

Within the IACS working group on Regional Assessments of Glacier Mass Change (RAGMAC) an intercomparsion experiment was set up to evaluate variations of volume change results between different processing chains and sensors. For this purpose, high resolution height change maps over two alpine glaciers, Aletschgletscher and Hintereisferner, were provided as independent validation data. Here we will target the influence of TanDEM-X bistatic scene selection for reporting height change rates and their difference to the provided validation data keeping a common processing chain. Ideally, two timely relevant, bistatic end of summer acquisitions from the same orbit and a suitable height of ambiguity should be utilized to calculate a DEM difference map from TanDEM-X data. In many cases however, the ideal set of criteria cannot be met and a scene selection from the TanDEM-X mission database has to be performed. SAR acquisitions from different seasons might enhance elevation biases due to signal penetration depending on the observed region. A combination of ascending and descending orbits will potentially increase the co-registration error and leads to an uneven sampling of the terrain due to different locations of shadow and layover caused by the steep terrain around many mountain glaciers. We will attempt to formulate a set of best-practices for TanDEM-X scene selection and future data acquisition ordering when targeting height change rate calculation for mountain glaciers.