| Land surface albedo(LSA)is an important parameter in surface radiation energy balance studies and an important driving factor in ground-air interactions.LSA reflects the level of reflection of the earth’s surface to solar radiation.Accurate estimation of land surface albedo is of great significance for global change,and medium and long-term weather prediction.At present,the successful launch of a series of domestic satellites has made significant contributions to Chinese quantitative remote sensing observation research.Compared with the progress of quantitative remote sensing observation research,the research on domestic satellite surface quantitative products is relatively lagging.In particular,the lack of research on surface albedo estimation methods and the consistency of albedo estimation results for domestic satellite data affects the service and promotion of domestic satellite data.Therefore,this study will develop a remote sensing estimation method for surface albedo based on the preprocessed multi-source domestic satellite data(GF-1 WFV,GF-6 WFV,CBERS-04WFI,HJ-1A/1B CCD),and explore the consistency among them.The contents of this paper include:(1)In this paper,a method for estimating the LSA of domestic satellites with various high spatial and temporal resolutions was proposed,and the surface albedo estimation model was constructed.The model estimates the LSA by directly establishing the statistical relationship between the top of atmosphere bidirectional reflectance in the narrowband and the LSA in the broadband.1)According to the spectral response function of MODIS and domestic satellites,their reflectance was calculated;The band conversion coefficients between MODIS and domestic satellite sensors were obtained from reflectance.2)The surface directional reflection characteristic parameters of MODIS were converted to the directional reflection characteristic parameters of each band of the domestic satellite sensors according to the band conversion coefficients.3)The 6S radiative transfer model was used to simulate the TOA reflectance under different geometric and atmospheric conditions.4)Establish a lookup grid and set up a three-level lookup table.Each grid established the linear regression relationship between the TOA reflectance and the broad band LSA of the domestic satellite sensors separately,and the regression coefficients of the linear regression were stored in the look-up table.5)Finally,the albedo of domestic satellites was estimated by using the look-up table,and the time series plot of the albedo of domestic satellites in 2019 was drawn.(2)Cross validation of albedo estimation results based on multi-source remote sensing data.Based on the field observation data of four meteorological sites in the study area,this paper adopted the verification method of site-by-site and of different land cover types.The albedo of sites was obtained from the ratio of upward radiation to downward radiation in the field observation data.Cross-validation of domestic satellite albedo was also performed based on Landsat data.The evaluation indexes included determination coefficient(R~2),root mean square error(RMSE)and bias.The validation results with field observation data showed that the R~2of all domestic satellites was 0.79,and RMSE was 0.04,among which the R~2of GF-1,CBERS-04 and HJ-1 exceeded 0.83,and that of GF-6 reached 0.50;which the RMSE of GF-1,GF-6 and CBERS-04 were less than 0.04,and that of GF-6 was 0.05.The validation results with Landsat albedo showed that the R~2of all domestic satellites was 0.87,and RMSE was 0.06,among which the R~2of GF-1,GF-6 and HJ-1 exceeded 0.93,and that of CBERS-04 reached0.57;which the RMSE of GF-1,GF-6 and HJ-1 were less than 0.04,and that of CBERS-04 was 0.07.The validation results showed that both have high accuracy,indicating that the estimation method used in this paper has high accuracy and can be applied to different domestic satellites.(3)Albedo consistency analysis of domestic satellites.The albedo results of domestic satellites with similar imaging dates were compared and analyzed their correlation.All satellites passed the 0.01 level of significance with each other,showing significant correlation.The strongest correlations were found between GF-1 and GF-6,and between CBERS-04 and HJ-1.All satellites showed strong correlation with each other except GF-1 which was moderately correlated with HJ-1.The correlation coefficient between GF-1 and GF-6,CBERS-04 and HJ-1 were the largest,reaching0.732,while the correlation coefficient between GF-1 and HJ-1 was the smallest,which was 0.584.The scatter plot results showed that the R~2range of domestic satellites was0.068-0.293 and the RMSE range was 0.037-0.079.The fit between GF-1,CBERS-04,and HJ-1 was good,while GF-6 fit poorly with the other three satellites,and the overall albedo value of GF-6 was higher than the others.The study also analyzed the possible error sources in the process of albedo estimation,and provided feasible solutions for each error source.In this study,using domestic satellite remote sensing images in 2018-2019,the albedo of the study area was estimated and verified,respectively.The time series of albedo in 2019 were established,and the consistency of albedo estimation between domestic satellites was analyzed.This study provides an effective method for albedo estimation of domestic satellites in the future,and facilitates the application and development of them. |
