An experiment to test satellite radar interferometry-observed geodetic ties to remotely monitor vertical land motion at tide gauges

摘要

The nature and linearity of vertical land motion (VLM) impacting the global sea level record from tide gauges is not well known, but remains of importance to understand long-term changes to sea level. Local surveys are required to directly measure VLM at tide gauges relative to a global reference frame, but this is limited by the lack of differential VLM measurements between tide gauges and continuously operating GPS (cGPS) stations that are not co-located, i.e., fixed to the tide gauge structure. We present results from an experiment using satellite radar interferometry (InSAR) scenes acquired from the TerraSAR-X satellite mission to test whether InSAR could replace repeat geodetic levelling as a 'geodetic tie' between cGPS stations and tide gauges. Comparisons are made among TerraSAR-X (TSX), cGPS and tide gauge minus altimetry VLM estimates for the Hillarys and Fremantle tide gauges (Perth, Western Australia), which are used as test sites for this method. The results suggest agreement between differential TSX and altimetry minus tide gauge VLM rates, but systematic offsets among the absolute/geocentric rates where the TSX is referenced to IGS08 at the PERT cGPS. The TerraSAR-X VLM at the Fremantle tide gauge for the period 7 October 2012 to 7 October 2017 is + 0.45 +/- 0.40 mm/yr (referenced to IGS08 at PERT cGPS), although this should be treated cautiously over this short period, and because VLM at Fremantle and Hillarys appears to be non-linear over time. We infer from this that the uncertainties in TerraSARX differential VLM rates are comparable to those from the highest quality repeat levelling, although the uncertainty approaches 1 mm/yr if the reference point uncertainty of the TSX and cGPS is considered when transformed to a terrestrial reference frame.