No abstract

We quantify GIA prediction uncertainties of 250 1D and 3D glacial isostatic adjustment (GIA) models through comparisons with deglacial relative sea-level (RSL) data from North America and rate of vertical land motion ( _ U) and gravity rate of change ( _ G) from GNSS and GRACE data, respectively. Spatially, the size of the RSL uncertainties varies across North America with the largest from Hudson Bay and near previous ice margins along the northern Atlantic and Pacific coasts, which suggests 3D viscosity structure in the lower mantle and laterally varying lithospheric thickness. Temporally, RSL uncertainties decrease from the Last Glacial Maximum to present except for west of Hudson Bay and the northeastern Pacific coast. The uncertainties of both these regions increase from 30 to 45 m between 15 and 11 ka BP, which may be due to the rapid decrease of surface loading at that time. Present-day _ U and _ G uncertainties are largest in southwestern Hudson Bay with magnitudes of 2.4 mm/year and 0.4 μGal/year, mainly due to the 3D viscosity structure in the lower mantle.

Coastal land is being lost worldwide at an alarming rate due to relative sea-level rise (RSLR) resulting from vertical land motion (VLM). This problem is understudied at a global scale, due to high spatial variability and difficulties reconciling VLM between regions. Here we provide self-consistent, high spatial resolution VLM observations derived from Interferometric Synthetic Aperture Radar for the 51 largest coastal cities, representing 22% of the global urban population. We show that peak subsidence rates are faster than current global mean sea-level rise rates and VLM contributions to RSLR are greater than IPCC projections in 90% and 53% of the cities respectively. Localized VLM worsens RSLR impacts on land and population in 73-75% of the cities, with Chittagong (Bangladesh), Yangon (Myanmar) and Jakarta (Indonesia) at greatest risk. With this dataset, accurate projections and comparisons of RSLR effects accounting for VLM are now possible for urban areas at a global scale. Sea-level rise resulting from climate change has rightly received substantial attention from researchers, practitioners and the public as an ongoing threat that needs to be addressed 1 . Yet lesser attention has been paid to land subsidence which can exceed tens of mm/year 2-4 , and increase local relative sea-level rise (RSLR) many times that of global mean sea-level rise of few mm/year alone 5,6 . Local RSLR, defined as sea-level rise relative to local land height, is what effectively matters for any coastal community. Furthermore, many coastal areas experiencing the fastest land subsidence are major cities built on flat, low elevation river deltas, exposing large populations and substantial economic value to the impacts of local RSLR 7,8 . Consequently, it is crucial to consider land subsidence when assessing coastal risks of RSLR 9,10 . Vertical land motion (VLM) -either subsidence (downward land motion) or uplift (upward land motion) -can be caused by several factors such as tectonics 11,12 , natural compaction of sediments 7,8 , groundwater, oil, and gas extraction 2,4 , reduced aggradation due to dams, levees, and loss of coastal vegetation 5 , and glacial isostatic adjustment 13,14 . Because these contributing factors vary significantly over a range of temporal and spatial scales, the contribution of VLM to RSLR has been difficult to assess on a global scale 15 . Many local-to regional-scale studies have mapped VLM at different coastal localities over different time