Quaternary International 95–96 (2002) 87–98 Geologically and geomorphologically constrained numerical model of Laurentide Ice Sheet inception and build-up Johan Kleman a, *, James Fastook b , Arjen P. Stroeven a a Department of Physical Geography and Quaternary Geology, Stockholm University, 10691 Stockholm, Sweden b Department of Computer Science, University of Maine, USA Abstract The locations, shapes and masses of evolving ice sheets during interglacial to glacial transitions remain elusive, but need to be clarified for identification of the climatic feedbacks that amplified low-amplitude insolation changes to become the major global climatic shift of the last glacial maximum. The best way to explore probable spatial patterns of ice sheet inception and build-up is through ice sheet modelling, and validating model output against geomorphological and geological data. We have modelled the 120–55 kyr BP evolution of the Laurentide Ice Sheet, focusing on the locations, outline, flow patterns, and basal thermal regimes of its precursor ice centres. For this, we employed a time-dependent mass-balance-driven finite element model, forced with GRIP ice core oxygen isotope values calibrated to temperature. The model achieves a good fit to relict ice flow systems in Keewatin, the Interior Plains, and the Hudson Bay lowland. Thus, our model indicates that the first coherent ice centre in North America was situated over Baffin Island and the Melville and Boothia Peninsulas, and expanded onto mainland Canada at 90–75 kyr BP. A previously unknown 200-km-long moraine zone near the Thlewiaza River, Keewatin, may demarcate the marginal position of this early Central Arctic Ice Sheet. Cordilleran and Keewatin ice merged, and the Hudson Bay lowland was inundated from the east (from Quebec), by 65 kyr BP. r 2002 Elsevier Science Ltd and INQUA. All rights reserved. 1. Introduction Our current understanding of the timing and magni- tude of global ice sheet build-up during the last interglacial–glacial transition is largely based on proxy data, and in particular on marine oxygen isotope records (Bond et al., 1993), coral reef records from tropical regions (Lambeck and Chappell, 2001), and on ice core records from Greenland and Antarctica (Johnsen et al., 1995). However, these proxy records cannot resolve the locations, configurations, and vo- lumes of individual ice sheets. Glacial geological and geomorphological field data of ice sheet inception and built-up yields, at best, ice sheet outlines and flow patterns. Numerical ice sheet models based on ice physics and forced with GCM (or otherwise)-derived mass-balance relationships, on the other hand, can yield ice sheet palaeotopography and volume for the time interval in question, i.e. facilitate full 3-D reconstruc- tions. However, model results diverge widely for individual ice sheets (e.g., Bintanja et al., 2002; Marshall, 2002). Hence, we invoke the need for a collaborative approach with strong geological/geomor- phological and numerical ice sheet modelling compo- nents. This is because ice sheet model reconstructions may be validated against field evidence (cf. Andrews and Mahaffy, 1976), even though the latter information is typically only available for restricted regions. We here present a plausible evolution of North American Ice Sheet centres during inception and built- up, who eventually coalesced to form the Laurentide Ice Sheet (LIS). We employ a mass-balance-driven finite element model for a time-dependent determination of the following ice sheet parameters; area, volume (Fig. 1), thickness (Fig. 2), surface topography, flow pattern (Fig. 3), and basal temperature through the time interval from 120 to 55 kyr BP. We intentionally focus on the 90–60 kyr BP time period, because the ice sheets had grown large enough to leave a geomorphological record in mainland Canada, i.e. the constraining data set of till lineations, striae, and till fabrics used to validate model performance. We sought spatial, directional and relative chronological convergence between model output and field evidence of relict ice flow patterns and ice marginal positions (i.e. older than the last glacial maximum). *Corresponding author. Tel.: +46-8-164-813; fax: +46-8-164-818. E-mail address: kleman@natgeo.su.se (J. Kleman). 1040-6182/02/$ - see front matter r 2002 Elsevier Science Ltd and INQUA. All rights reserved. PII:S1040-6182(02)00030-7