Estimation of Long-term Gridded Cloud Radiative Kernel and Radiative Effects Based on Cloud Fraction

Abstract. The surface shortwave cloud radiative effect (CRE) plays a critical role in modulating the Earth's energy balance and climate change. However, accurately quantifying the CRE remains challenging due to significant uncertainties in downwelling surface shortwave radiation (DSSR) and cloud parameter estimates, especially in the Arctic. This paper introduces a novel approach that enhances the accuracy of CRE estimation by constructing a computationally efficient, long-term gridded surface cloud fraction radiative kernels (GCF-CRKs) and integrating refined DSSR estimates and a high-precision cloud fraction (CF). By leveraging the correlation between the top-of-atmosphere (TOA) shortwave radiative parameters and surface radiation, combined with high-precision fused CF datasets from multiple satellite sources, we construct a CF-dependent model to refine DSSR estimates. Based on this model, we construct GCF-CRKs using the CF as the sole perturbation parameter to isolate the CF CRE. Our results indicate that this method significantly improves the accuracy of DSSR estimation under partially cloudy conditions (0<CF<100 %), aligning more closely with ground-based observations. In Arctic-wide validation experiments, the root mean square error (RMSE) was decreased by approximately 2.5 Wm^-2, and the bias was reduced by 1.23 Wm^-2, which was an improvement of 8.7 % (reduction of RMSE) against the CERES-EBAF. The even greater improvements were achieved at stations in Greenland (RMSE reduced by 4.53 Wm^-2 and a bias reduced by ~6.89 Wm^-2, with an accuracy improved about 11.1%). The GCF-CRKs exhibit similar signs and patterns and enhanced stability compared to existing kernels. The sensitivity analysis results reveal that seasonal and interannual variations introduce GCF-CRK uncertainties of approximately 1 Wm^-2 %^-1 and 0.1 Wm^-2 %^-1, respectively, while spatial variations within the same latitude range can cause CRK uncertainties of 0.2–1.2 Wm^-2 %^-1. These uncertainties can result in CRE biases ranging from 5 to 50 Wm^-2, which demonstrates the limitations of existing methods that utilize short-term, small-area parameter data to produce global CRKs. Using these GCF-CRKs, we estimated the spatiotemporal properties of the surface shortwave CRE in the Arctic over a 21-year period (2000–2020), and the trend result indicates that despite the increasing influence of the CF on the Arctic DSSR, the smaller magnitude and interannual trend of the annual average surface shortwave CRE suggest that previous studies may have overestimated the magnitude and rate of the cooling effect of clouds on the Arctic DSSR by up to 4 Wm^-2 and 0.5 Wm^-2 per decade, particularly in Greenland. This study provides a more accurate and efficient assessment of the CRE, and the results underscore the need for more effective measures to mitigate the impact of Arctic amplification on the surface radiative energy balance, which is crucial for understanding and addressing regional and global climate change. The GCF-CRKs can be freely available to the public at https://doi.org/10.5281/zenodo.13907217 (Liu, 2024).