Variation and comparison of cloud cover in MODIS and four reanalysis datasets of ERA-interim, ERA5, MERRA-2 and NCEP

Cloud plays an important role in climate change and radiative forcing. Atmospheric reanalysis datasets provide an economical and efficient way to explore the worldwide cloud variation. This study compares the monthly total cloud cover (TCC) and high cloud cover (HCC) of the four commonly used atmospheric reanalysis datasets, including European Centre for Medium-Range Weather Forecasts (ECMWF) Interim Reanalysis (ERA-Interim), ECMWF's Fifth-generation Reanalysis (ERA5), Modern-Era Retrospective Analysis for Research and Applications, Version 2 (MERRA-2) and National Centers for Environmental Prediction (NCEP: only TCC), with MODIS re-trievals during 2001-2018. The spatial and temporal distributions of TCC for the four reanalyses are generally consistent with MODIS results, while showing better consistency over land than over ocean. The largest TCC discrepancy between MODIS and the reanalysis appears over the polar regions. Among the four reanalysis datasets, ERA5 presents the highest correlation and the smallest mean bias compared with MODIS, while NCEP shows the lowest and the largest. Although all the three reanalyses (ERA-Interim, ERA5 and MERRA-2) generally give an overestimation of HCC compared with MODIS, the ERA-Interim shows the closest to MODIS. The largest difference of HCCs between MODIS and the reanalysis is located over the Tibet Plateau. To the annual change of global means, ERA-Interim and MERRA-2 appear a significant increasing trend of TCC over both land and ocean. ERA5 and MERRA-2 show a significant decreasing trend of HCC over ocean, and MODIS gives a reduce over land. Furthermore, During the ENSO events, a positive anomaly of TCC, which is positively correlated with the ENSO intensity over the eastern equatorial Pacific, is captured by all the four reanalyses, among which ERA5 and ERA -Interim appear more distinct.