Observations show that increased Arctic cloud cover in the spring is linked with sea ice decline. As the atmosphere and sea ice can influence each other, which one plays the leading role in spring remains unclear. Here we demonstrate, through observational data diagnosis and numerical modeling, that there is active coupling between the atmosphere and sea ice in early spring. Sea ice melting and thus the presence of more open water lead to stronger evaporation and promote cloud formation that increases downward longwave flux, leading to even more ice melt. Spring clouds are a driving force in the disappearance of sea ice and displacing the mechanism of atmosphere‐sea ice coupling from April to June. These results suggest the need to accurately model interactions of Arctic clouds and radiation in Earth System Models in order to improve projections of the future of the Arctic.

Reanalyses have proven to be convenient tools for studying the Arctic climate system, but their uncertainties should first be identified. In this study, five reanalyses (JRA-55, 20CRv2c, CFSR, ERA-Interim, and MERRA-2) are compared with NASA CERES–MODIS (CM)-derived cloud fractions (CFs), cloud water paths (CWPs), top-of-atmosphere (TOA) and surface longwave (LW) and shortwave (SW) radiative fluxes over the Arctic (70°–90°N) over the period of 2000–12, and CloudSat–CALIPSO (CC)-derived CFs from 2006 to 2010. The monthly mean CFs in all reanalyses except JRA-55 are close to or slightly higher than the CC-derived CFs from May to September. However, wintertime CF cannot be confidently evaluated until instrument simulators are implemented in reanalysis products. The comparison between CM and CC CFs indicates that CM-derived CFs are reliable in summer but not in winter. Although the reanalysis CWPs follow the general seasonal variations of CM CWPs, their annual means are only half or even less than the CM-retrieved CWPs (126 g m−2). The annual mean differences in TOA and surface SW and LW fluxes between CERES EBAF and reanalyses are less than 6 W m−2 for TOA radiative fluxes and 16 W m−2 for surface radiative fluxes. All reanalyses show positive biases along the northern and eastern coasts of Greenland as a result of model elevation biases or possible CM clear-sky retrieval issues. The correlations between the reanalyses and CERES satellite retrievals indicate that all five reanalyses estimate radiative fluxes better than cloud properties, and MERRA-2 and JRA-55 exhibit comparatively higher correlations for Arctic cloud and radiation properties.

The decline of sea ice in the Arctic region is a critical indicator of rapid global warming and can also influence the feedback processes in the Arctic, so the prediction of sea ice extent and thickness plays an important role in climate modeling and prediction. This paper uses machine learning methods to predict the sea ice extent, and by adjusting the methods and factors, which include the climate variables, the past sea ice extent, and the simple linear-regression-simulated sea ice extent, then we found the best combination to give the result with the highest R² score. We noticed that with longer periods of past sea ice extent data and shorter periods of climate data, the results appeared to be better. This might be related to the difference in climate and ocean memory. The sub-region sea ice extent prediction shows that the regions with whole-year ice cover are easier to predict and that those regions with sudden weather changes and significant seasonal variability appear to have lower R² scores in the sea ice extent prediction.

Low stratiform clouds have profound impacts on the hydrological cycle and the Earth’s radiation budget. However, realistic simulation of low clouds in climate models presents a major challenge. Here we employ the newly retrieved cloud and drizzle microphysical properties to improve the autoconversion and accretion parameterizations in a microphysical scheme. We find that the new autoconversion (accretion) rate contributes 14% lower (greater) to total drizzle water content than the original scheme near the cloud top. Compared to satellite results, the simulated cloud liquid water path (LWP) and shortwave cloud radiative effect using the original scheme in a climate model agree well on global average but with large regional differences. Simulations using the updated scheme show a 7.3% decrease in the light rain frequency, and a 10% increase in LWP. The updated microphysics scheme alleviates the long‐lasting problem in most climate models, that is “too frequent and too light precipitation.”