Estimation Of Net Surface Shortwave Radiation From Chinese Geostationary Meteorological Satellite FengYun-2D(FY-2D) Data

As a key driving factor of the surface energy budget and balance,surface short wave net radiation（NSSR）affects obviously the regional and global climate change.So far,many researchers have proposed a variety of methods to estimate the NSSR.However,there is less research on the estimation of NSSR using Chinese geostationary meteorological satellite（FY-2D）data.Regarding the method proposed by Tang et al.（2006）as a reference,based on FY-2D data simulated by the atmospheric radiative transfer model（MODTRAN4）,a method to estimate NSSR is developed.Firstly,a linear regression model is established between the top of the atmosphere（TOA）broadband albedo（r）and the narrowband reflectivity（pi）based on the data simulated with MODTRAN 4.2 under clear and cloud sky,respectively.Secondly,the relationship of surface absorption coefficient（a_s）and broadband albedo（r）is set up under clear and cloud sky,respectively,by dividing the land type into land,sea,snow and ice,respectively,and NSSR can be calculated using the above formulae.Last,validations are performed.The simulation results show that the root mean square error（RMSE）between the estimated and real r is less than 0.0151 for all conditions under clear sky,and the RMSE of estimated and real NSSR are 25.7432 W/m²,10.1038 W/m²,22.5303 W/m² for land,sea,snow and ice,respectively.Under cloudy sky,whether cumulus,altostratus,stratus or three kinds of hybrid cloud,the RMSE between the estimated and real r is less than 0.0116,and the RMSE of estimated and simulated NSSR is the most if the same coefficients are used to calculate NSSR for the three kinds of clouds,which is 19.3346 W/m²、11.8353 W/m²、20.7291 W/m² for land,sea,snow and ice,respectively.Considering the errors of the input parameters of atmospheric water vapor content（w）,the top of atmospheric narrowband reflectivity（ρ₁）,solar zenith angle（SZA）and viewing zenith angle（VZA）will affect the estimated accuracy,we analysis the sensitivity of these factors under clear sky.The results show that the error of pi affect on estimation of NSSR greatest,the error of w affect on estimation of NS SR minimum,land surface affected by these factors is smallest among three types of land,and these factors affect on estimation of NSSR are within acceptable limits.At the same time,comparison of field measurement in TaiYuan and Changwu respectively with estimation using corresponding FY-2D data acquired from January to march,2012 under cloud-free and cloud days is performed respectively,and result shows that the RMSE between the estimated and actual NSSRs is 48.56 W/m² with mean error of-2.23 W/m² in Taiyuan,the RMSE between the estimated and actual NSSRs is 73.93 W/m² with mean error of 13.62 W/m² in Changwu.The RMSE between the estimated and actual NSSRs is 88.99 W/m² with mean error of 45.77 W/m² in Taiyuan under cloud days.This shows that the proposed method is feasible and has practical application value.