Validation and comparison of surface shortwave and longwave radiation products over the three poles

Abstract

Global warming has currently become a great concern to the international community, among which the three poles (the Arctic, Antarctic, and Qinghai-Tibet Plateau) are the most serious. In this paper, in order to improve the understanding of the matter and energy cycle in the three poles and even the world, eleven shortwave products, (namely, CERES-SYN, ERA5, MERRA-2, NCEP-CFSR, JRA-55, GLDAS, BESS_Rad, MCD18A1, ISCCP-HXG-SSR, GLASS and APP-x), and seven longwave products, (CERES-SYN, ERA5, MERRA-2, NCEP-CFSR, JRA-55, GLDAS, and APP-x) are evaluated and inter-compared in terms of accuracy. During the assessment, the ground measurements collected from four independent ground observation networks (BSRN, CEOP, TPDC, and NMC) are used as reference, and the being compared products are aggregated to the same spatial and temporal scales to make them comparable. To better examine their performance, the eleven radiation products are comprehensively compared in multiple spatial (original scale and 1°×1°) and temporal scales (1-hourly, 3-hourly, daily, and monthly means, and instantaneous). The results show that in the three poles, CERES-SYN and ERA5 show overall better accuracy, at daily, 1°×1° resolutions. The (r)bias and (r)RMSE are less than (3%)5 W/m2 and (23%)40 W/m2 for SWDR, and less than (3%)7 W/m2 and (15%)25 W/m2 for LWDR, respectively, over the polar regions (the Arctic and Antarctic), which are generally better than that of the Qinghai-Tibet Plateau for most products. The remote sensing products and reanalysis products have their own advantages and disadvantages at different regions. In addition, with the spatio-temporal resolution decreasing, the accuracy of radiation products will gradually increase, except for the products of MCD18A1, ISCCP-HXG-SSR, NCEP-CFSR and GLDAS.