ALBERT

Description: Space geodetic studies of both present-day surface mass trend (PDMT) and GIA are significantly hindered by the duality of signatures in various data. Current GIA models are largely built on global forward modeling approaches conducted in the 1990s with subsequent piecemeal improvements often without knowing PDMT. With a general lack of uncertainty assessments, they are not accurate enough to match modern space geodesy precisions, and contain possible ambiguities and large errors. We carry out a joint dynamic inversion of multiple data sets of different space geodetic techniques and historical relative sea level records to solve for PDMT, ice history and Earth rheology simultaneously. The dynamic GIA forward modeling is based on gravitationally and topographically self-consistent sea-level equation solver SELEN 4.0. The deglaciation process features 759 global equal-area, icosahedron-shaped, spherical ice pixels with an approximate radius of 1.34 degrees, 13 2-kyr Heaviside steps, and a 4-layer Earth model. Loose a priori PDMT and GIA models are used based on our earlier kinematic inversions and ICE-6G/VM5A respectively. An innovative finite difference method is developed to overcome the enormous computational cost of evaluating partial derivatives with respect to ice thickness and Earth rheology parameters in the context of solving integral equations. Our results confirm many PDMT and GIA signatures estimated in previous kinematic inversions. But two disjoint minimums of equal size with distinct lower mantle viscosities are seen to both satisfy the data combination. Significant coherent ice history deviations from ICE-6G are also found.