Remote Sensing Estimation Method Of Surface Albedo From FY-3C MERSI Data

Surface albedo is a key surface characteristic parameter in the study of radiation and energy balance,which can reflect the reflection ability of the surface to solar radiation and is also a key parameter in the calculation of net surface radiation.It has been widely used in surface energy balance,medium and long-term weather forecasting and global climate change.The specific work of this paper includes:(1)MERSI data preprocessing: MERSI data detecting radiation calibration,abnormal banding,cloud and GLT geometric correction,such as pretreatment,formed a relatively complete and effective MERSI data preprocessing algorithm,data preprocessing to get the overall quality is very high,basic did not reduce the quality of the raw data,for after the albedo inversion has laid a solid foundation.Data preprocessing results directly affect the quality of albedo,in the process of data and processing of abnormal banding detection and cloud detection is studied,abnormal banding detection adopted band correlation method,and compared with mathematical morphology method to do the validation,is adopted in this paper according to the results of abnormal banding detection method based on spectrum correlation can be more effective for detecting abnormal stripe removal,including anomaly detection accuracy is as high as 98.1%,abnormal all columns can be accurately detected.Cloud detection research mainly refer to MODIS cloud detection algorithm for cloud detection algorithm suitable for MERSI data,this article for cloud detection is divided into two parts,land and sea,although some problems still exist in the cloud detection,such as high reflectivity of the earth's surface and cloud failed to fully distinguish,land part of the cloud to remove clean test but on the whole,most of the cloud can be identified,especially can identify the sea area of the cloud.Radiation calibration can be directly carried out by using the calibration coefficient in MERSI data set,and geometric correction can be carried out by using GLT geometric correction method,which has higher accuracy.The overall and processing results can be used as the basic data for inversion of surface albedo.(2)Direct inversion algorithm: The overall idea of this method is to directly establish the regression relationship between atmospheric directional reflectivity and surface wide-band albedo,so as to directly estimate surface wide-band albedo based on TOA directional reflectivity from a single observation Angle,and generate surface albedo data set with high temporal resolution.First,spectral resampling.Then,obtain the band conversion coefficient of MODIS and MERSI through band conversion.Last,convert MODIS albedo lookup table into MERSI lookup table based on band conversion coefficient,so as to establish the multivariate linear regression relationship between atmospheric topside albedo and surface wide band albedo.In this paper using the classification(divided into vegetation,soil,land,sea water and sea ice,snow and ice)and grid points Angle(the sun zenith Angle/observation 2°,5° Angle between interval),set up to the top of the atmosphere direction reflectivity and surface albedo wide band of multivariate linear regression relation.RSI atmospheric topside reflectivity and surface wide band albedo was generated.(3)Validation of surface albedo estimations: Based on this algorithm generates the northeast Asian region of high time resolution(1 day),1 km spatial resolution of the surface albedo data sets,10 in the United States(including 6 kinds of IGBP terrain types)FLUXNET flux site observation data and surface albedo MODIS products(MCD43A3),to generate the FY-3C MERSI surface albedo data set a preliminary comparison and verification.In this paper,site by site verification and ground object type verification were carried out,and the selected evaluation indexes were R2,RMSE and BIAS.Existing authentication day resolution of the inversion results show that the algorithm to get the surface albedo products in three evaluation indexes overall than MODIS surface albedo product(MCD43A3)is better,it shows that the inversion of this algorithm is of a higher accuracy of the surface albedo,can better describe the change of the surface albedo,suitable for use in a regional or global climate change research.This algorithm can be applied to the surface albedo products with high precision and high temporal resolution.