Updating global energy balance based on the latest observations and reanalyses

The energy balance at the top of the atmosphere (TOA) and at the surface of the earth plays an essential role in the Earth's climate change. The Clouds and Earth Radiant Energy System (CERES) can directly measure the energy radiation at the TOA, while the surface energy balance needs to be estimated by satellite retrievals and atmospheric reanalysis. Some key fluxes need to be adjusted artificially to close the surface energy balance, particularly the surface latent and sensible heat fluxes. Existing studies have used the Global Precipitation Climatology Project (GPCP) dataset to estimate latent heat flux and have adjusted it artificially to balance the surface energy. Based on high-precision ground-based and satellite observations, this study evaluates the ability of the new generation of atmospheric reanalyses to quantify the global energy balance and provides the optimal estimation of the global, terrestrial, and oceanic radiation fluxes at the TOA and the surface during the period 2015-2020. Additionally, the Global Precipitation Measurement (GPM) products are combined to provide the best estimates of the surface latent heat flux. The global mean TOA incident solar radiation from the new generation of reanalyses is in good agreement with that from CERES. However, due to the addition of atmospheric tide analysis, the Climate Forecast System Reanalysis (CFSR) has a harmonic oscillation pattern that appears near the equator; the errors cancel out after taking the global mean, but the results still have a significant overestimation of 6 W m(-2) over land. On the global scale, it is found that the Modern-Era Retrospective Analysis for Research and Applications V2 (MERRA2) overestimates the TOA reflected solar radiation by 8 W m(-2), and the Japanese 55-year Reanalysis (JRA55) overestimates the TOA outgoing longwave radiation by 10 W m(-2), which is most significant over the tropical ocean. The European Centre for Medium-Range Weather Forecasts Reanalysis of Atmosphere Version 5 (ERA5) can reproduce the TOA energy imbalance well (1.1 W m(-2)), while the estimates of the other three reanalyses are 5.4 W m(-2) (CFSR), -10.7 W m(-2) (JRA55), and -4.1 W m(-2) (MERRA2). All of the TOA radiation fluxes of the reanalyses over the ocean are consistent with those at the global scale, and the ERA5 results are still the best. The best estimates of the surface incident solar radiation and downward longwave radiation from CERES and ERA5 are 187 and 342 W m(-2), respectively. Additionally, due to atmospheric tidal analysis, CFSR seriously overestimates the surface incident solar radiation, JRA55 and MERRA2 yield overestimates over land, and MERRA2 seriously underestimates the surface downward longwave radiation. Compared with the latest GPM products, the existing precipitation estimates based on GPCP underestimate global precipitation by 8.4%, mainly due to the substantial underestimation of precipitation over the ocean, which helps to close the global surface energy balance. Based on the GPM and the estimation of runoff into the ocean, we conclude that the optimal estimates of surface latent heat flux over the globe, over the land, and over the ocean are 83, 41, and 101 W m(-2), respectively. The global and oceanic latent heat fluxes of JRA55 are overestimated by 10 and 13 W m(-2), respectively, while the other three reanalyses are within reasonable ranges. According to the evaluation results, this study provides the optimal estimation of each component of the surface energy balance, which can reach surface energy closure without artificial adjustment.

Automated summary

This study evaluates the performance of atmospheric reanalysis data in quantifying the global energy balance and provides optimal estimates of radiation fluxes at the top of the atmosphere and the surface. It identifies issues with some reanalysis data, such as overestimation or underestimation of solar and longwave radiation, as well as discrepancies in latent heat flux. By combining precipitation data with runoff estimates, the study improves the estimation of surface energy balance.