Patagonian Ice Sheet extent as an indicator of the regional climate regime at the Last Glacial Maximum

During the Last Glacial Maximum (LGM, 23 to 19 thousand years ago), the Patagonian Ice Sheet (PIS) expanded along the Andes between ~ 38 °S to 55 °S. Existing paleoclimatic and paleoglacial evidence, especially that derived from glacial landforms, seems to indicate that the timing of maximum ice sheet expansions in the Southern and Northern Hemispheres was not synchronous. Moreover, significant uncertainties still exist in the onset of regional deglaciation and its major drivers. In this context, a combination of ice sheet modelling, glacial geochronology and paleoclimate reconstructions can provide important insights into the former PIS geometry and its contribution to the sea level low during the LGM. It can also help us infer likely paleoclimate scenarios and climate models that capture regional climate responses to global change in the most realistic manner.

Here we present an ensemble of numerical ice sheet simulations of the PIS at the LGM to constrain an envelope of probable atmospheric conditions derived from a range of model-based climate forcing products from the phases 3 and 4 of the Paleoclimate Modelling Intercomparison Project (PMIP). The resulting ensemble is then used as a guideline to identify sectors of the PIS where a significant disagreement between the field evidence and modelling results is obtained, highlighting a strong dependence of the PIS geometry on the uncertainties in near-surface air temperature forcing. We find that all ensemble members consistently fail to reproduce the ice sheet extent towards the northern part of Patagonia within the explored model parameter space. At the same time, the modelled PIS expands beyond its southeastern reconstructed boundary. Our analysis of the ice sheet’s mass budget seems to indicate that these discrepancies between the modelled and reconstructed PIS extents arise from poorly resolved topographic features within the global climate models and the general lack of observational data on ice thickness distribution during the LGM. We conclude that INM-CM4-8 and MPI-ESM1-2-LR produce the most realistic climate forcing across Patagonia at the LGM. It should be kept in mind that this analysis is based only on the evaluation of modelled ice sheet extents against geological evidence, as observational data on the former ice sheet thickness are still lacking. Nevertheless, our analysis suggests that the quality of the regional model-based climate reconstructions is directly linked to the horizontal resolution that must be capable of resolving topographic features of the Andes and ideally of the PIS itself.