Assessing the portion of historical Greenland surface mass balance change attributable to anthropogenic forcing and its uncertainties

Greenland's contribution to global mean sea level exhibits decadal variability, driven by interannual surface mass balance (SMB) changes. In this study, we attribute historical Greenland SMB changes to radiative forcings using the Community Earth System Model version 2 Large Ensemble and its single-forcing Large Ensemble simulations (CESM2-LE and CESM2-SFLE), which enables separation of impacts from greenhouse gases and aerosols. We quantify the contribution of radiative forcings to Greenland SMB changes by estimating univariate and multivariate detection and attribution scaling factors through Bayesian total least squares regression implemented via Markov Chain Monte Carlo (MCMC). The MCMC formulation allows us to quantify the uncertainty of the scaling factors using prior knowledge from observation-based simulations and reconstructions, as well as CESM2-LE and CESM2-SFLE. Our results indicate that historical Greenland SMB changes can be attributed to anthropogenic forcings, including anthropogenic aerosols, which affect decadal scale variability superimposed on the greenhouse gas-driven long-term trend. However, CESM2 tends to underestimate the relative contribution of each individual forcing to observed historical Greenland SMB changes. To explore potential reasons for this underestimation, we test a few hypotheses, including the role of internal variability. Our analysis demonstrates that internal variability plays only a minor role in the underestimation of the forced Greenland SMB changes due to individual forcings. Additionally, we find that Greenland runoff changes, rather than precipitation changes, explain both the SMB changes and the underestimation of attributable portions to individual forcings. Our findings emphasize the confounding role of aerosol forcing on the historical SMB trajectory but also highlight outstanding questions regarding the ability of climate models to correctly parse such influences. We will discuss the implications of these issues and steps to address them.