Validation And Spatio-Temporal Comparison Of Longwave And Shortwave Radiation Products Over The "Three Poles"

The earth’s surface radiation budget mainly consists of shortwave radiation(0.3～3μm)and longwave radiation(4～100μm).Radiation,as the main driving force of atmospheric motion,directly determines the global carbon,water and other material cycles,and is the basis for the study of global climate change.Due to the particularity of natural conditions and geographical location of the "three poles" region(Arctic,Antarctic and Qinghai-Tibet Plateau),the longwave and shortwave radiation in this region has a significant impact on global climate change and surface radiation balance.However,the current systematic research on the longwave and shortwave radiation over the "three poles" region is still weak,and people lack the understanding of the spatio-temporal variation of the radiation component in the "three poles" region,and there is a lack of longwave and shortwave radiation research algorithms with high spatial and temporal resolution and high accuracy.Therefore,it is particularly important to study the longwave and shortwave radiation over the "three-pole" region and to propose a high-precision algorithm for longwave and shortwave radiation algorithms.In order to deepen the understanding of the longwave and shortwave radiation over the "three poles" region,the existing eleven longwave and shortwave radiation products are comprehensively evaluated and inter-compared in terms of accuracy over the "three poles" region.During the assessment,the thirty-five ground measurements collected from four independent ground observation networks are used as reference to comprehensively compare radiation products at multiple spatial scales(original scale and 1°×1° spatial resolution)and multiple time scales(3-hour average,daily average,monthly average and instantaneous).In addition,the existing long wave downward radiation algorithm is improved to solve the problems of low representation of training data,low accuracy,lack of verification of global applicability,etc.The main conclusions of this paper are as follows:(1)A comprehensive assessment of eleven radiation products was carried out using the observation data of ground stations over the "three poles" region.The results show that the radiation products conclusions are similar on the 3 hour-ly average,daily average and monthly average.Taking the verification of daily mean,1°×1° spatial resolutions as an example,in the polar regions,CERES-SYN,ERA5 and NCEP-CFSR products have better accuracy for SWDR,(r)bias and(r)RMSE are less than(3%)7W/m~2 and(28%)40 W/m~2,respectively,while the accuracy of other reanalysis and remote sensing products are relatively poor;For LWDR,except for MERRA-2 and JRA-5 reanalysis products with poor accuracy,CERES-SYN and other reanalysis products have higher accuracy,with(r)bias and(r)RMSE generally less than(4%)6.5W/m~2 and(14%)25 W/m~2;While the accuracy of the longwave and shortwave net radiation is generally poor,and the r RMSE is generally greater than 50%.In the Qinghai-Tibet Plateau,except for BESS＿Rad,MCD18A1 and ISCCP-HXG-SSR products,the accuracy of SWDR is significantly better than the polar regions,and the absolute value of(r)bias is less than(2.5%)5.5 W/m~2,and the accuracy of most radiation products is generally lower than that of the polar 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.At the same time,the spatio-temporal analysis shows that although the precision of the above radiation products is different in time and space series,the overall change trend is relatively consistent.(2)In this paper,the training database of the existing longwave downward radiation algorithm is improved and the parameterization coefficient of the algorithm is updated,which has wider applicability than the original algorithm.Globally verified,the new algorithm has high accuracy(bias less than 3 W/m~2,RMSE less than 25 W/m~2).At the same time,compared with the original algorithm,the new algorithm also solves the problem of underestimation in high altitude area and overestimation in high latitude area.