Quantifying the Uncertainties of Reanalyzed Arctic Cloud-radiation Properties Using Satellite-surface Observations

Abstract:

Cloud and radiation processes are very important issues in Arctic climate system. Reanalyses have proved to be the essential tools to study extreme weather and climate events, especially in data-sparse region like the Arctic. Before using reanalyses products, their strengths and uncertainties should be identified. In this study, five recent reanalyses (JRA55, 20CR V2c, CFSR, ERA-Interim and MERRA) are compared with NASA CERES satellite observations with respect of cloud fraction (CF), top-of-atmosphere (TOA) and surface longwave (LW)/shortwave (SW) radiation fluxes during the period of 03/2000-02/2012 over the Arctic (70-90°N). 20CR V2c, CFSR, ERA-Interim and MERRA overestimate CFs, particularly during the cold season, with the positive biases of annual means from +9.6% (MERRA) to +22.9% (20CR V2c). Only JRA55 can represent its overall seasonal variation and spatial distribution but with large negative biases (nearly -15%). All reanalyses can well capture the seasonal trend of TOA SW/LW upwelling fluxes. However, in all-sky condition, all of them show positive biases of TOA SW upwelling flux along northern and eastern coasts in Greenland during the warm season (JJA). There is a good agreement between reanalyses and observation in seasonal cycle of net TOA cloud radiative effects (CRE), which are calculated by TOA SW/LW fluxes. The spatial distributions of net TOA CRE in warm season show that only JRA55 and ERA-Interim are relatively consistent with their reanalyzed CFs. As for the surface radiation, the satellite-derived results were firstly validated by Baseline Surface Radiation Network (BSRN) ground-based observations. It illustrates that average biases of satellite retrievals are +9.85 W/m² for surface downward SW flux and +0.39 W/m² for downward LW flux in warm season within the Arctic. The seasonal variation of SW/LW fluxes can be well represented by four of five reanalyses except MERRA. Reanalyzed surface downward SW flux in JRA55, CFSR and ERA-Interim are relatively consistent with their CF results among these reanalyses. However, the biases in TOA and surface radiation fluxes cannot only explained by biased CFs in some of reanalyses.