During the Last Glacial Maximum, ice sheets covered large areas in northern latitudes and global temperatures were significantly lower than today. But few direct estimates exist of the volume of the ice sheets, or the timing and rates of change during their advance and retreat. Here we analyse four distinct sediment facies in the shallow, tectonically stable Bonaparte Gulf, Australia--each of which is characteristic of a distinct range in sea level--to estimate the maximum volume of land-based ice during the last glaciation and the timing of the initial melting phase. We use faunal assemblages and preservation status of the sediments to distinguish open marine, shallow marine, marginal marine and brackish conditions, and estimate the timing and the mass of the ice sheets using radiocarbon dating and glacio-hydro-isostatic modelling. Our results indicate that from at least 22,000 to 19,000 (calendar) years before present, land-based ice volume was at its maximum, exceeding today's grounded ice sheets by 52.5 x 10(6) km. A rapid decrease in ice volume by about 10% within a few hundred years terminated the Last Glacial Maximum at 19,000 +/- 250 years.

No abstract

Northwestern Europe remains a key region for testing models of glacial isostasy because of the good geological record of crustal response to the glacial unloading since the time of the Last Glacial Maximum. Models for this rebound and associated sealevel change require a detailed knowledge of the ice-sheet geometry, including the ice thickness through time. Existing ice-sheet reconstructions are strongly model-dependent, and inversions of sea-level data for the mantle response may be a function of the model assumptions. Thus inverse solutions for the sea-level data are sought that include both ice-and earth-model parameters as unknowns. Sea-level data from Fennoscandia, the North Sea, the British Isles and the Atlantic and English Channel coasts have been evaluated and incorporated into the solutions. The starting ice sheet for Fennoscandia is based on a reconstruction of a model by Denton & Hughes (1981) that is characterized by quasi-parabolic cross-sections and symmetry about the load centre. Both global (northwestern Europe as a whole) and regional (subsets of the data) solutions have been made for earth-model parameters and ice-height scaling parameters.The key results are as follows. (1) The response of the upper mantle to the changing ice and water loads is spatially relatively homogenous across Scandinavia, the North Sea and the British Isles. (2) This response can be adequately modelled by an effective elastic lithosphere of thickness 65-85 km and by an effective upper-mantle viscosity (from the base of the lithosphere to the 670 km depth seismic discontinuity) of about 3-4×1020 Pa s. The effective lower-mantle viscosity is at least an order of magnitude greater. (3) The ice thickness over Scandinavia at the time of maximum glaciation was only about 2000 m, much less than the 3400 m assumed in the Denton & Hughes model. ( 4) The ice profiles are asymmetric about the centre of the ice sheet with those over the western part being consistent with quasi-parabolic functions whereas the ice heights over the eastern and southern regions increase much more slowly with distance inwards from the ice margin.

UK Efficacy and Mechanism Evaluation Programme and National Institutes of Health.

No abstract