Uncertainty in East Antarctic firn thickness constrained using a model ensemble approach

Mass balance assessments of the East Antarctic ice sheet are highly sensitive to changes in firn thickness resulting from variability in firn compaction rates and surface mass fluxes (snowfall, sublimation, melt). To better constrain uncertainty in firn thickness and in the underlying processes, we develop a model-based ensemble of firn evolution scenarios over 1992-2017. We combine statistical emulation of nine firn-densification models, climatic output from three regional climate models and different assumptions about surface snow density to generate a comprehensive set of 54 model scenarios. The ensemble agrees that firn thickness changes in the interior are minor, but there are pronounced thickening and thinning patterns in coastal areas.  At basin level, model uncertainty in firn thickness change ranges between 0.2–1.0 cm yr^-1 (15–300%). Statistical analysis of the ensemble uncertainty demonstrates that climatic forcing is the primary contributor of model spread on firn thickness estimates. However, in basins characterised by warmer temperatures, high snowfall or increasing snowfall, the contributions of firn compaction and surface snow density can account for up to 46 and 28% of the spread, respectively.

By comparing the ensemble scenarios with satellite measurements of elevation changes over the same 1992-2017 period, we find that these estimates are consistent over a majority of basins. Nonetheless, we identify several basins where model estimates of firn thickness change do not match altimetry measurements. These discrepancies can be explained by different causes: (1) the model ensemble may fail to represent the real firn thickness change over our period of interest, (2) the uncertainty range associated with the altimetry data may not capture the true signal and (3) a component of the elevation change signal may be related to ice dynamical imbalance. As such, our analysis serves to highlight specific areas where further focus on potential sources of errors in model and altimetry results is needed in order to better constrain mass balance assessments in East Antarctica.