Searching for laterally heterogeneous models of glacial isostatic adjustment with the ICE-6G_C ice history model

Abstract

Most models of Glacial Isostatic Adjustment (GIA) assume that the Earth is laterally homogeneous. However, seismic and geological observations clearly show that the Earth's mantle is laterally heterogeneous. Previous studies of GIA with lateral heterogeneity mostly focused on its effects or sensitivities on GIA predictions, and it is not clear to what extent can lateral heterogeneity solve the misfits between GIA predictions and observations. Our aim is to search for the best laterally heterogeneous GIA models that can simultaneously fit the global relative sea-level (RSL) data, the peak uplift rates (u-dot from GNSS) and peak gravity-rate-of-change (g-dot from the GRACE satellite mission) in Laurentia and Fennoscandia. However, the search is dependent on the ice and viscosity model inputs - the latter depends on the background viscosity and the seismic tomography models used. In the thesis, the ICE-6G_C ice model, with Bunge & Grand's seismic tomography model and background viscosity models close to VM5 will be assumed.

A Coupled Laplace-Finite Element Method is used to compute gravitationally self-consistent sea level change with time dependent coastlines and rotational feedback in addition to changes in deformation, gravity and the state of stress.

Several laterally heterogeneous mantle viscosity models are found to fit the global sea level data better than laterally homogeneous models. Two of these laterally heterogeneous mantle viscosity models also fit the peak g-dot and u-dot rates observed in Laurentia simultaneously. However, even with the introduction of lateral heterogeneity, no model that is able to fit the present-day g-dot and u-dot data in Fennoscandia has been found.

Then the effects of laterally heterogeneous lithosphere, sub-lithospheric and asthenospheric properties are studied and we confirm that they can affect the predicted global RSL, present-day g-dot and u-dot in Laurentia and Fennoscandia. In addition, incorporating the laterally heterogeneous lithosphere can improve the fit to global RSL, but the values of g-dot and u-dot in Laurentia may decrease slightly but not significant enough to affect the fit to the observed data. Our results prefer an elastic lithosphere that has maximum thickness of 140 km under continental cratons but reduces to 60 km underneath the oceans. In addition, the results preferred depth of the asthenospheric bottom is around 190-200 km with asthenospheric viscosity around $10^{20}$ Pa s. Next, we show that the best laterally heterogeneous mantle model when combined with the best laterally heterogenous lithospheric model give the best fit to global RSL and peak g-dot and u-dot in Laurentia simultaneously.

Next, we allow the heterogeneity under Fennoscandia to be different from that under Laurentia and show that this can help to further improve the fit to the observed data. Also the sensitivity of our results to different ice history model, seismic tomography model and background viscosity profile are investigated. Finally, we confirm that the ice history of ICE-6G_C in Fennoscandia and Barents Sea needs some modifications.