3D MANTLE VISCOSITY STRUCTURE IN GLACIAL ISOSTATIC ADJUSTMENT MODELS RESOLVES DISCREPANCIES IN MARINE ISOTOPE STAGE (MIS) 5A AND 5C GLOBAL MEAN SEA LEVEL PREDICTIONS (Invited Presentation)

Global geophysical models of glacial isostatic adjustment (GIA) investigate geographic patterns in high stand elevations reconstructed from paleo-sea level indicators to refine global mean sea level (GMSL) during past warm periods. Yet, analyses of distinct geographic gradients in MIS 5a/5c peak high stand elevations along the North America Atlantic (Potter and Lambeck, 2003) and Pacific (Simms et al., 2016) coastlines have yielded discrepant GMSL estimates for these substages (e.g., MIS 5a peak high stand ∼28 m versus ∼15 m below present, respectively). Here we explore the suggestion of Creveling et al. (2017) that accounting for lateral variations in mantle viscosity structure beneath North American in GIA models will simultaneously fit high stand elevations along both North American coastlines.

To test this suggestion, we construct a 3D viscoelastic Earth model as follows: (1) we adopt lateral variations in lithospheric thickness (LT) and mantle viscosity derived by Richards et al. (2018) and Hoggard et al. (2020) based on the SL2013sv shear-wave velocity model of North American upper mantle viscosity structure (Schaeffer and Lebedev, 2013; 2014) and S40RTS in the lower mantle (Ritsema et al., 2011); (2) we superimpose these variations on a spherically averaged (‘1-D’) mantle viscosity profile favored in 1-D GIA analyses of MIS 5a/5c high stand data from US Atlantic coast and Caribbean (Potter and Lambeck, 2003; Creveling et al., 2017). This 1-D model is comprised of upper and lower mantle viscosities of 0.5×10²¹ and 5.0×10²¹ Pa s, respectively, and an 80 km LT. We coupled this Earth model to the Pico et al. (2017) ice history from MIS 5e to the Late Holocene, which is characterized by MIS 5a and 5c peak GMSL of ∼ -14 m and -9 m, respectively, relative to present day. Finally, we generate gravitationally self-consistent sea-level predictions for all North American and Caribbean field sites based upon this 3-D Earth model/ice history pairing. We find that the resulting GIA predictions simultaneously fit observed MIS 5a/5c high stand elevations along both coasts of North America and the Caribbean. The analysis refines existing constraints on MIS 5a/5c peak GMSL values (and, thus, ice volumes), and represents the first demonstration that 3D GIA modeling can reconcile large compilations of regional sea-level indicators.