Tide model accuracy in the Amundsen Sea, Antarctica, from radar interferometry observations of ice shelf motion

This study assesses the accuracy of tide model predictions in the Amundsen Sea sector of West Antarctica. Tide model accuracy in this region is poorly constrained, yet tide models contribute to simulations of ocean heat transfer and to the removal of tidal signals from satellite observations of ice shelves. We use two satellite‐based interferometric synthetic aperture radar (InSAR) methods to measure the tidal motion of the Dotson Ice Shelf at multiple epochs: a single‐difference technique that measures tidal displacement and a double‐difference technique that measures changes in tidal displacement. We use these observations to evaluate predictions from three tide models (TPXO7.1, CATS2008a_opt, and FES2004). All three models perform comparably well, exhibiting root‐mean‐square deviations from the observations of ∼9 cm (single‐difference technique) and ∼10 cm (double‐difference technique). Care should be taken in generalizing these error statistics because (1) the Dotson Ice Shelf experiences relatively small semidiurnal tides and (2) our observations are not sensitive to all tidal constituents. An error analysis of our InSAR‐based methods indicates measurement errors of 7 and 4 cm for the single‐ and double‐difference techniques, respectively. A model‐based correction for the effect of fluctuations in atmospheric pressure yields an ∼6% improvement in the agreement between tide model predictions and observations. This study suggests that tide model accuracy in the Amundsen Sea is comparable to other Antarctic regions where tide models are better constrained. These methods can be used to evaluate tide models in other remote Antarctic waters.