Separating Long-Term and Short-Term Mass Changes of Antarctic Ice Drainage Basins: A Coupled State Space Analysis of Satellite Observations and Model Products

Abstract

Satellite gravimetry and altimetry measurements record gravity and elevation changes, respectively, which are useful for determining mass and volume change of the Antarctic Ice Sheet. Common methods employ products from regional climate modeling and firn modeling to aid interpretation and to link volume changes to mass changes. Estimating deterministic parameters over defined time periods is a conventional way to evaluate those changes. To overcome limitations of deterministic analyses with respect to time-variable signals, we have developed a state-space model framework. Therein, we jointly evaluate four mass and volume data sets by coupling of temporal signal variations. We identify long-term signals of ice drainage basins that are observed by the satellite gravimetry mission GRACE and several satellite altimetry missions from April 2002 until August 2016. The degree to which we can separate long-term and short-term variations strongly depends on the similarity of the mass and volume change time series. For the drainage system of the Pine Island Glacier (West Antarctica), our results show noticeable variations of the long-term trend with an acceleration of the contribution of ice dynamics to the mass balance from −11 ± 8 to −58 ± 8 Gt a−1. Our results in Dronning Maud Land (East Antarctica) show a positive long-term contribution to the mass balance at almost a constant rate. The presented approach can fit time-variable changes without artificial selection of periods of interest. Furthermore, because we only enforce common long-term time variations between mass and volume data, differences in mean trend rates help to uncover model discrepancies.