In search of an ice history that is consistent with composite rheology in Glacial Isostatic Adjustment modelling

Abstract

Global reconstructions of the Earth's Quaternary ice sheet history such as ICE-6G assume that mantle rheology is linear and Earth properties are laterally homogeneous. However, these assumptions are unrealistic because high temperature/pressure creep experiments show that both linear and non-linear creep laws operate simultaneously in the mantle. Also surface geology and seismic tomography show that mantle properties vary laterally. This study uses the Coupled Laplace-Finite Element Method to find an ice model that is consistent with an Earth model with composite rheology where both linear and non-linear creep laws operate simultaneously in the mantle, and the effective viscosity is laterally heterogeneous.We use the ICE-6G ice model as initial load on a composite rheology Earth model and iteratively improve both the ice and Earth model parameters until they fit observations of Glacial Isostatic Adjustment or GIA (relative sea level and land uplift rate) well simultaneously. We thereby focus on North America and northern Europe. Our best combination model, called ICE-C(M_C), also fits gravity rate-of-change data in both areas. This finding removes the obstacle in using composite rheology in GIA studies since previous work found it impossible for composite rheology to fit both relative sea level and land uplift rate data simultaneously. It shows that composite rheology is clearly an alternative to linear rheology in future GIA studies.The ICE-C model is at the last glacial maximum (LGM) position from 26 ka BP (thousand years before present) to 19 kaBP, which implies that composite rheology also supports the timing of LGM proposed by Clark et al. (2009). The model is found to be significantly thicker in the center of rebound in Fennoscandia than ICE-6G from the LGM to about 8 kaBP, but significantly thinner around the Fennoscandian coastline from LGM to about 12 kaBP. These changes in the lower boundary conditions in northern Europe may affect the reconstruction of past climate there. Regional differences in ice thickness between ICE-C and ICE-6G are also found in Laurentia. During LGM there are negligible differences in total ice volume between ICE-C and ICE-6G, thus it remains questionable if a composite rheology may help in solving the missing ice problem. ICE-C is also consistent with meltwater pulses 1A and 1B. Mantle viscosities and creep parameters of Earth model M_C are consistent with previous findings from GIA modeling and microphysics experiments.