How Does Cloud Emissivity Feedback Affect Present and Future Arctic Warming?

The Arctic is warming at a rate higher than the global average. Notwithstanding the differences between models and observations, particularly those based on clouds, the physical linkage between cloud process representations and Arctic climate projections remains uncertain. Here, we investigated the simulated and observationally estimated relationships between cloud–radiation interactions and Arctic warming using 30 CMIP6 models, satellite retrievals, and reanalysis data. The underestimated cloud liquid fraction in the models yielded insufficient cloud longwave emissivity and underestimated cloud radiative surface warming. In response to future warming, excessive variations in the ice-to-liquid cloud phase would result in an overestimated increase in cloud emissivity. This process enhances cloud radiative surface warming (and thus Arctic warming) until cloud emissivity attains an asymptotic value of one (blackbody). This implies that the feedback process termed “cloud emissivity positive feedback” in this paper is time-dependent. This, in turn, emphasizes the importance of the cloud phase representation that is consistent with observations of the present climate. Our results explain why the wide inter-model spreads in projected future Arctic warming can be attributed to cloud bias. Thus, these provide a better understanding of uncertain factors in atmosphere–ocean–cryosphere interactions.