Greenland's committed sea-level contribution calibrated by altimetry observations

Abstract:

Sea-level rise projections over the coming century rely on ice-sheet models to simulate ice sheet response to climate warming. However, calibrating ice sheet models so that they reproduce the observed changes of the recent past remains challenging and an active area of research. This led the recent Ice Sheet Model Intercomparison (ISMIP6) to adopt an anomaly-based approach that quantifies future ice sheet contribution to sea-level rise, neglecting the committed response already "locked-in" to the ice sheets due to recent forcing, also known as committed sea-level rise. One approach used to tune numerical models is Bayesian calibration, where an ensemble of model simulations is created by varying parameters and parameterizations with a prescribed prior distribution. The model simulations are split into a hindcast, which overlaps with observations, and a forecast, which simulates future behavior. Discrepancy between model output and observations are used to weight each ensemble member, yielding a calibrated posterior distribution of simulations. Here, we calibrate committed sea-level rise projections of the Greenland Ice Sheet using surface elevation change measurements from NASA's spaceborne and airborne laser altimetry record, as well as digital elevation models. We explore various metrics for measuring model-observation discrepancy and their effect on the calibrated posterior distribution of sea-level rise in 2100.