. Isolating the influence of cloud radiative feedbacks on Arctic amplification through cloud-locking

Abstract
Satellite observations show increased cloud cover during non-summer months associated with Arctic sea ice loss, but no cloud response to summer sea ice loss (Morrison et al. 2018a; Kay et al. 2016). These observed non-summer cloud increases can increase surface-based Arctic amplification by generating a positive longwave cloud radiative feedback. Some climate models can reproduce observed present-day cloud-sea ice relationships. For example, CESM1 captures the physical mechanisms controlling seasonal changes in the cloud response to sea ice (Morrison et al. 2018b). Yet, it remains challenging to diagnose cloud influence on Arctic amplification in climate models because diagnostic techniques used to quantify cloud feedbacks struggle to account for nonlinearity among radiative feedbacks. Here, we utilize a technique called "cloud-locking" to isolate the influence of cloud radiative feedbacks on Arctic amplification within CESM1. Specifically, we disable Arctic cloud radiative feedbacks by prescribing one year of cloud fields recurrently in the model's radiative transfer calculations from 70N poleward. By analyzing surface temperatures in response to a CO2 doubling both with and without cloud radiative feedbacks in the Arctic, we separate the influence of cloud radiative feedbacks on Arctic amplification. Because CESM replicates observed Arctic cloud-sea ice relationships but has insufficient opaque cloud, we anticipate a small influence of cloud radiative feedbacks on Arctic amplification. The novelty of this experiment, though, lies in the fact that we are effectively isolating the role of cloud radiative feedbacks in a state-of-the-art model while accounting for the nonlinearity in the feedbacks between cloud radiative effects and the Arctic climate system.