Ice mass change and its feedback on solid earth dynamics in the Antarctic peninsula

Rapid regional climate warming in the Antarctic Peninsula (AP) has led to several major ice shelves retreating, and eventually collapsing, since the 1970s. In response, feeding glaciers have exhibited rapid acceleration and thinning, contributing to sea-level rise. This ice mass unloading induces a solid Earth response, which can be measured by geodetic observations. Observed rates of three-dimensional solid Earth deformations contain contributions due to both present and past ice mass variations (i.e. since Last Glacial Maximum, LGM) and horizontal plate tectonics. The solid Earth viscoelastic adjustments due to ice-ocean loading changes are known as Glacial Isostatic Adjustment (GIA). Accurate knowledge of GIA is essential for correcting gravity-based estimates of present-day ice-mass change. Observing solid Earth deformation caused by present-day ice mass change can be used to constrain the GIA response by inferring the Earth's viscoelastic properties. A few studies based on seismic or geodetic datasets have inferred Earth's rheological properties in the AP but the results are not yet conclusive. This thesis seeks to improve understanding of the solid Earth rheology in the AP by using spatiotemporally extended, three-dimensional geodetic observations and viscoelastic modeling. GPS observations up to 2018 are used to constrain the Earth rheology in the region around the former Larsen A and B ice shelves. I make use of interferometric synthetic aperture radar (InSAR) observations to further understand the ice mass change and solid Earth rheology in the northern Antarctic Peninsula (NAP), the first such application of InSAR in Antarctica. Finally, I extend the study region further south, to the northern Marguerite Bay (NMB) region, to determine ice mass change over 2002-2018 and explain non-linear deformation observed in GPS records using viscoelastic modeling. Previous work has shown that GPS measurements of bedrock uplift in the NAP can only be explained by a viscoelastic response to post-2002 ice surface unloading, with upper-mantle viscosities of ~ <2√v=10\\(^{18}\\) Pa s and a wide range of elastic lithosphere thicknesses. This thesis shows that since around 2011, the GPS uplift rates have reduced which is in accord with recent estimates of much-reduced ice mass loss in this region. The extended GPS uplift analysis and viscoelastic modeling confirms earlier estimates of low upper-mantle viscosities in this region but shows little sensitivity to variations in modeled lithospheric thickness. Along with GPS-derived bedrock uplift time series, the horizontal GPS components are also investigated for the first time. The horizontal GPS displacements are directed towards the south-west, in accord with the known and ongoing ice mass loss in the eastern Peninsula. The east coordinate component is shown to be sensitive to ice mass change, and modeling of it confirms that this region is underlain by the upper mantle viscosities suggested by the GPS uplift rates. In order to consider the horizontal components, uncertain plate motion must be removed from the time series, and so four test scenarios of plate motion were tested. The results from this experiment indicate that the ITRF2014 plate motion model accurately describes the Palmer (PALM) station motion and suggests that any potential plate model error or post-LGM GIA signal has a magnitude of ~ ¬± 0.5 mm/year for this region. I expanded the GPS datasets from a small number of stations to include spatially-extensive InSAR observations by the Sentinel-1A C-band dataset, consisting of nearly 85 and 84 acquisitions during the 2014.9-2017.8 period over the Larsen A and B regions, respectively. In the region of Larsen-B, large relative line-of-sight (LOS) displacements are observed at outlet glaciers of low elevation where ice unloading is high. InSAR also indicates that mass loss around the southern part of the Larsen-A region is higher relative to the northern part. Comparing these InSAR data and far-field GPS-based results to updated viscoelastic modeling for the Larsen-B region refines the understanding of lithospheric thickness, demonstrating a poor fit to a thin lithosphere. InSAR shows a good agreement for lithospheric thicknesses in the range of ~85-130 km with the upper-mantle viscosities preferred from comparison with the GPS time series. This is an early and potential application of InSAR to Antarctic deformation and can be further improved with better knowledge of tropospheric water vapour in this region. This thesis also investigates the ice mass changes from ~2002 around the NMB region, about ~350 km south of the Larsen-B, and subsequent solid Earth deformation. The mass balance estimation over this region suggests that the ice mass loss reduced around the Rothera research station since ~2012 and the Muller Ice Shelf since ~2009 compared to 2004-2012 and 2002- 2009, respectively. GPS measurements of bedrock uplift in NMB show time-varying rates of uplift varying between ~2.2 and 7.0 mm/year over 2002-2018. A comparison between GPS and modeled viscoelastic deformation up to 2015 suggests an upper mantle viscosity of ~0.1- 80√v=10\\(^{18}\\) Pa s but allows a wide range of effective elastic lithosphere thickness for NMB. The precise solid Earth properties of the AP remain to be conclusively determined, but our investigation sets an approximate range for the effective lithospheric thickness and upper mantle viscosity, which will help to understand more complex features of the solid Earth in this region in the future.