Estimation of downward surface shortwave radiation from Himawari-8 atmospheric products

Abstract

Downward Surface Shortwave Radiation (DSSR) estimated from satellite measurements is crucial in climate change study and clean energy applications. The Advanced Himawari Imager (AHI) onboard the new generation geostationary satellite Himawari-8 provides an unprecedented opportunity for the near real-time estimation of DSSR, with a spatial resolution of 5 km and a temporal resolution of 10 min over the full disk regions. To meet the requirements of fast and accurate estimation of DSSR from Himawari-8, this study proposed a Look-Up Table (LUT) method to estimate the DSSR from Himawari-8/AHI level 2 (L2) atmospheric products. We first investigate the sensitivities of DSSR to solar geometry (solar zenith angle), atmosphere conditions (aerosol optical depth, cloud optical depth, and cloud effective radius), and surface condition (surface albedo) basing on the atmospheric radiative transfer model. Then, LUTs for clear and cloudy skies are generated based on the sensitivity results. Finally, the DSSR is estimated with the inputs of Himawari-8 L2 aerosol and cloud products released by JAXA on the basis of LUTs previously created. As an experiment, DSSR results are estimated using our algorithm at 02:00 UTC on April 1, 2016 and compared with the JAXA Himawari-8 L2 DSSR product. The comparison shows that our DSSR estimates are consistent with the operational DSSR results over the full disk regions. To further validate our DSSR estimates, we compare our results and the operational DSSR results with ground-based measurements at Yonsei site (land) and 0n_165e site (sea) in April, July, October, and December 2016. The correlation coefficients (R) derived from ground measurements and these two DSSR results are larger than 0.88 for all types of sky conditions. With the scattering properties of non-spherical (hexagon) ice cloud particles included, the biases of our DSSR estimates in the validation with ground measurements at two sites are lower than the operational DSSR results. This study developed an LUT-based method to estimate DSSR with inputs of Himawari-8 L2 atmospheric products (including aerosol and cloud products, and other auxiliary data such as solar zenith angle in L1 product). The estimated DSSR was validated against both land and sea sites of ground observed DSSR, with RMSEs of 94.13 Wm-2, 62.92 Wm-2, and 110.60 Wm-2 for all types of sky conditions at Yonsei, and RMSEs of 123.86 Wm-2, 105.33 Wm-2 and 151.44 Wm-2 for all types of sky conditions at sea site 0n_165e. Furthermore, the correlation coefficients (R) of our DSSR estimation from the land and sea sites were greater than 0.88 for all sky conditions. These validation results suggested that our DSSR estimation with Himawari-8 atmospheric products works well and can thus be further used in land surface radiation budget research and solar energy application after improvements on the current algorithm.