Research On An Efficient Radiative Transfer Model For The Forward Modeling Of Remote Sensing

In this paper,a ?-6-stream spherical harmonic expansion adding method(?-6SDA)for solar radiative transfer is developed,and a radiative transfer model suitable for forward modeling of remote sensing is established based on ?-6SDA.?-6SDA can not only be used to calculate the radiation flux but also can be used to calculate the azimuthally dependent radiation intensity.In the radiation flux calculation,the accuracy of ?-6SDA is similar to that of ?-6-stream discrete ordinates radiative transfer(?-6DISORT).The results of ?-6SDA are more accurate than that of ?-4-stream spherical harmonic expansion adding method(?-4SDA)and ?-4-stream discrete ordinates radiative Transfer(?-4DISORT).The computational efficiency of ?-6SDA is three orders of magnitude faster than ?-6DISORT.The truncated-plus-single-scattering(TMS)correction is used in the calculation of the azimuthally dependent radiation intensity.In the Cloud C1 and Henyey-Greenstein phase function cases,the relative errors of ?-6SDA are less than 2% and 5%,respectively,which are smaller than those of ?-4DISORT and ?-6DISORT.The computing speed of ?-6SDA is 3-4 orders of magnitude faster than ?-6DISORT,which shows the potential of its application in satellite remote sensing research.Besides,a radiative transfer model is established based on ?-6SDA and applied to the simulation of AHI onboard Himawari-8 satellite observation to test its performance.Typhoon Hato and Maria are chosen as typical cases of cloudy weather to be studied.The maximum Root Mean Squared Error(RMSE)between the simulated results of the established model and a strict radiative transfer model(e.g.,?-256DISORT)is about 0.016.The computing speed of this model is about 4 orders of magnitude faster than the rigorous model.The absolute error between simulated reflectance of the established model and observation of Himawari-8 / AHI is less than 0.1 and the simulation results can reflect the shape and characteristics of Typhoon Hato and Maria.The RMSE of water cloud simulation in different channels is within 0.05.The error of the ice cloud is slightly larger,but the maximum RMSE of ice cloud is only about 0.16.The error is mainly due to the uncertainty of cloud optical properties and other atmospheric products.The model itself has high accuracy and efficiency.Therefore,this model can be applied to climate models and satellite remote sensing.