New monthly maps of accumulation over the Greenland Ice Sheet

The Greenland Ice Sheet (GrIS) has become the single largest contributor to present day sea-level rise, with mass loss driven by changes in Surface Mass Balance (SMB). As the largest component of SMB, snow accumulation is critical to monitor as Arctic warming continues at an accelerated rate. Snowfall patterns across GrIS are influenced by a complex interaction of many interdependent climate variables, leading to high inter-annual spatial variability. As a result, regional climate models (RCMs) often fail to adequately capture this variability and carry substantial uncertainties, leading to biased estimations of ice mass loss. Here, we present a novel method to bias-adjust RCM precipitation output with in-situ SMB records from the SUMup dataset (2024 release), including over two million data points from radar, ice-core, snow pit and stake measurements. RCM output data is first decomposed into Empirical Orthogonal Functions (EOFs), reflecting different modes of spatial variability, and Principal Components (PCs), capturing temporal fluctuations correlating to various climate indices. The SUMup in-situ measurements are used to derive a set of coefficients to scale the model mean climatology and each EOF and PC through least-squares optimisation. We provide monthly bias-adjusted accumulation maps for HIRHAM5-ERA5 output between 1960-2023 and CARRA between 1991-2023, highlighting regional biases in the models through time. 

Preliminary mean bias maps for HIRHAM5 show that the model underestimates accumulation in the south and interiors of the ice sheet by 20-80% or 30-90 mm/year, while the west and east margins of the accumulation zone are overestimated by 20-60% or 30-150 mm/year. In the winter and spring, the model tends to underestimate accumulation overall by 50-100 mm/year, while the reverse is true for the summer and autumn, when accumulation is mostly overestimated, reaching up to 200 mm/year in the north west.