Numerical modelling of postglacial rebound predics a spatially variable pattern of sea-level change. Yet, many previous models predicting Late Pleistocene sea-level change either neglect the load of water added to the oceans, or assume a spatially uniform time-dependent load over the whole ocean. This is a poor approximation of the load near the former ice sheet, since sea-level change varies geographically. In addition, the coastline goemetry changes through time.
A spectral technique has been used to solve the sea-level equation on an earth model with axisymmetric distribution of ice and oceans and also for a realistic ice-ocean configuration. The axisymmetric model is useful for demonstrating the fundamental physics involved in sea-level changes caused by changes in surface loads. In particular, it can be seen that the elastic and gravitational response to the load is spread over a great distance, whereas the viscous response is more localized. A realistic ice-ocean configuration has been used to demonstrate that, neglect of the melt-water load underestimates the amount of uplift in Fennoscandia and other regions where rebound occurs in the ocean [as previously noted by Wu & Peltier (1983) in Hudson Bay], and that movement of the coastlines and spatially varying water loads must both be included in calculations for the North Sea and other near field regions with shallow seas. Sea-level calculations including these effects and careful analysis of the sea-level and ice-melting record are required to determine modifications to the viscosity profile and ice model inferred by previous inaccurate models of the ocean load in the near field.