Using reconstructed ice streams to calibrate a coupled climate-ice-sheet model of the North American Ice Sheet Complex during the Last Glacial Maximum

Coupled climate-ice-sheet modelling provides critical insights into the mechanisms underlying ice-sheet-climate feedback. These processes have strong implications for past and future climate change events, yet modelling efforts remain constrained by uncertainties in key model parameters. To address this limitation, we rely on comparisons between model outputs and available records of past ice sheets. Historically, this involved matching simulated ice sheets to reconstructed extent and volume derived from a range of geomorphological and sea level change data. Although these metrics are useful to validate ice sheet geometry and volume, they only provide limited information on ice sheet dynamics. New methods, which compare the footprint of reconstructed and simulated palaeo-ice streams, offer promising ways to incorporate a dynamical dimension into model calibration (Ely et al., 2024, Journal of Quaternary Science). 

In this project, we catalogue the distinct dynamical configurations observed in an ensemble of coupled climate-ice-sheet simulations of the Last Glacial Maximum (LGM, 21,000 years ago). This ensemble includes 124 equilibrium simulations generated using the coupled atmosphere-ice-sheet model FAMOUS-BISICLES, with variation applied to 12 model parameters representing ice dynamics, albedo and climate feedbacks (Patterson et al., 2025, EGUsphere). The ice sheet dynamics not only assess the model’s ability to replicate the LGM reconstructions of the Laurentide ice streams (Margold et al. 2018, Quaternary Science Reviews), but they also inform the sensitivity of the simulated ice sheets to climate forcing.  

Plausible simulations of the North American ice sheets in terms of volume and extent can be obtained across various regions of the parameters space, resulting in significant discrepancies in potential ice streaming patterns. Surface Mass Balance (SMB) is the main factor behind these changes in dynamical configurations: simulations with low accumulation tend to produce less numerous and intense ice streams, whereas high accumulation is associated with more vigorous ice streaming. In addition, the parametrisation of the ice dynamics influences the location and consistency of the ice streams, as well as the ability of the ice sheet to respond to climate change events. 

We find that simulations with relatively high SMB and ice dynamics parameters that enable fast-flowing and well-defined ice streams best match estimates of Last Glacial Maximum North American ice sheet extent, volume and ice stream location. Conversely, high friction coefficients and porous subglacial till, or low resolution of the ice sheet margins and the bedrock topography, result in ice stream patterns that are inconsistent with reconstructions and less responsive to climate forcing. This work demonstrates the relevance of comparison between reconstructions of past ice streams and model simulations to provide strong constraints on dynamical ice sheet models and ice sheet sensitivity to climate changes.