Abstract
Surface mass balance (SMB) of the Antarctic Ice Sheet (AIS) is an important contributor to global sea level change. Past climates provide an opportunity to evaluate model performance outside the range of recent observed climate variability. We look to the Last Millennium (850–1850 CE) as a period of relative climate stability to understand what processes control natural variability in SMB, distinct from anthropogenic warming. With evidence for large regional differences in climate trends from ice core proxy records, paleo-simulations need to be validated over long timescales to assess if they capture those regional variations. The drivers for such regional variations during the Last Millennium and present day remain uncertain, demonstrating the need for a regionally focused study. Here, we evaluate model performance by comparing available Paleoclimate Modelling Intercomparison Project (PMIP) past1000 models and the CESM Last Millennium Ensemble (CESM-LME) to four sets of Last Millennium Antarctic proxy-based reconstructions that are most relevant to the SMB: snow accumulation, surface air temperature (SAT), sea surface temperature (SST) and Niño 3.4 index, using a multi-parameter scoring method. Our results show that, overall, PMIP past1000 models reasonably capture SATs estimated in the proxy record, but show poor skill with respect to reconstructed regional snow accumulation means, trends and variability and the Niño 3.4 index. Models show some skill but a slight cold bias in simulating Southern Ocean SST. The overall best-scoring PMIP past1000 models for regional climate features of Antarctica and the Southern Ocean are the CESM-LME mean and CSIRO-Mk3L-1-2. CESM-LME predicts higher SMB by 2100.