Construction of late Pleistocene Laurentide Ice History on earth with composite rheology

Abstract

A good ice thickness history model is essential in the study of Glacial Isostatic Adjustment (GIA), its effects on coastal engineering, water resource management, fault stability and intraplate earthquakes, monitor global climate change, etc... Ice history models can be constructed based on glaciology and climate data only, but Peltier mainly used GIA observations and simple ice physics to construct global models ICE-4G, 5G & 6G where the main uncertainty is the ice thickness in Antarctica and western Laurentide during the last glacial maximum. One should note that most of the ice models constructed this way are based on the assumption that mantle rheology is linear and that rheology varies in the radial direction only. However, surface geology and seismic tomography show that Earth properties also vary strongly in the lateral direction. Moreover, high temperature creep experiments on mantle rocks show that mantle flow is better described by composite rheology since both diffusion (linear) and dislocation (nonlinear) creep operate in the mantle at the same time. The aim of our study is to construct global ice history models that are consistent with composite rheology and lateral heterogeneity. Thus we use the Coupled Poisson-Finite Element method to model GIA in a spherical, self-gravitating viscoelastic Earth with composite rheology and lateral heterogeneity. We shall follow the approach of Peltier and use GIA observations and simple ice physics as constraints to our ice model. The limitation of using sea level data is that they only lie near the coast and thus there is little constraint on ice thickness inland. To overcome this, we will use gravity rate-of-change data from GRACE with the effect of hydrology accurately removed using GPS observations (rather than GIA models which introduce large uncertainties). However, these data only give the current-day rate-of-change, which is more than 8,000 years after the end of deglaciation. To further constrain our ice model, we also use river incision data which records tilting of the land since the formation of the river after local ice melted. We will also use the paleo-stress orientations induced from postglacial faults formed near the end of deglaciation to further constrain ice thickness changes. Some preliminary results will be presented.