Modelling glacial isostatic adjustment with composite rheology

Abstract

Modelling glacial isostatic adjustment (GIA) can be employed to infer mantle rheologic structure and reconstruct ice history. The former is important in understanding mantle convection process while the later can provide an insight in understanding sea level rise, glacier stability and climate change. Most previous work on GIA modelling is based on linear rheology and assumes that the Earth is laterally homogeneous. However, the Earth is laterally heterogeneous, and high temperature and high pressure creep experiments of mantle rocks show that both linear rheology and non-linear rheology can operate simultaneously in the mantle. The work presented in this thesis employs a more realistic rheology to model GIA, i.e. composite rheology where linear rheology and non-linear rheology operate simultaneously and the effective viscosity is laterally heterogeneous. The first piece of work in this thesis is modelling the thermal effects of viscous heating induced by GIA on Earth models with composite rheology, linear rheology and non-linear rheology. It has been argued that viscous heating induced by GIA can trigger transient volcanism, change mantle properties, or affect the stability of ice-sheets with surface heating. Our results show that in general, composite rheology and non-linear rheology can produce higher viscous heating than linear rheology. Although the perturbation to surface heat flux induced by viscous heating is not negligible, the perturbation to mantle temperature field is so small that viscous heating cannot induce volcanism, affect mantle properties or even ice stability. The second piece of work is on searching for the best ice history model that is consistent with composite rheology. Although composite rheology has been employed to model GIA, no ice model consistent with composite rheology has been found to be able to fit both observed RSL and land uplift rates in Laurentia and Fennoscandia simultaneously. We find that with a layered composite rheology, it is possible to construct an ice history model that can fit the observed RSL, uplift rates and gravity -rates-of-change simultaneously. The best ice model and rheological model obtained from the search are termed as ICE-C and M_C respectively. ICE-C differs from the currently popular ICE-6G in both North America and Europe, and contains more ice mass from 17 thousand years before present (kBP) to 6 kBP and the parameters in M_C is consistent with laboratory findings. The last piece of work is to study the important question of whether there is any stress interaction between GIA-induced stress and tectonic stress from mantle convection. Wu (2001) has shown that for non-linear rheology, such interaction between GIA-induced stress and stress induced by mantle convection exists. With composite rheology in a spherical self-gravitating earth, we find that there is also an interaction and the interaction has an influence on the deformation and the geoid. The magnitude of interaction depends on the level of tectonic stress. The sensitivity of RSL and gravity rate-of-change to the interaction shows that tectonic stress in Eurasia and North America plates cannot be greater than 10 MPa.