Remote Retrieval Of Atmospheric CO₂ And CH₄ Using GOSAT

The study of global carbon cycle and climate change is urgent because of the intensified greenhouse effect.Global carbon cycle research and global climate change research puts forward higher requirements for the accuracy of greenhouse gas CO₂ and CH₄.Satellite-based remote sensing detection of greenhouse gases can provide long-term and continuous observations at the global scale and has become the primary observational model for monitoring global greenhouse gas changes.With the development of science and technology,the spatial resolution and time resolution of the sensor for observing greenhouse gases are improving.On the other hand,the requirement for accurate retrieval of greenhouse gases is increasing.Therefore,the study of high-precision retrieval of greenhouse gases is a hot topic now.In this paper,atmospheric radiation transmission simulation and retrieval method of atmospheric CO₂ and CH₄ are studied.To understand the impact of others parameters on retrievals of CO₂ and CH₄,it is necessary to analyze the sensitivity of these parameters.The effect of parameter on the simulation has been analyzed by changing the initial value of a parameter while fixation values of other variables.The main error of retrieval of CO₂ and CH₄ is from albedo,aerosol and water vapor.Therefore,the sensitivity of these three parameters are analyzed.The concept of normalized sensitivity is introduced in sensitivity analysis to study the relative changes between the each inversion bands.In the aerosol sensitivity analysis,the seasonal parameters,the boundary layer aerosol type and the boundary layer visibility are studied.Results shows that the boundary layer visibility has the most influence on the rate of radiance change.The change rate of the radiance of the visibility 5km is about five times that of the change of 23km.The boundary layer aerosol type has the secondary influence on the rate of radiance change.The change rate of the radiance of urban aerosol type is about two times that of the change of country aerosol type.The seasonal parameters has the least influence on the rate of radiance change.Spring and summer type is about 0.2 percent larger than autumn and winter type.In the normalized sensitivity analysis,the boundary layer visibility has the greatest impact on the difference band.The boundary layer visibility has the greatest impact on the difference band.The boundary layer aerosol type has the least impact on the difference band.The normalized sensitivity at six surface albedos between 0.05 to 0.4 has been analyzed.Results show that the normalized sensitivities at these surface albedos are with one consistent.The impact of water vapor profile increasing from 5%to 10%shows that most bands are not sensitive to the change of water vapor profile.After the removing of sensitive bands,the impact of water vapor profile are negligible.In the study of sensitivity analysis,the influence of aerosol,surface albedo and water vapor profile are consistent between CO₂ and CH₄.The channels for retrieval are selected based on the normalized sensitivity of aerosols and surface albedo.We obtain 91 high-sensitivity channels and 285low-sensitivity channels for CO₂ between 1560 nm and 1590 nm,and 41high-sensitivity channels and 641 low-sensitivity channels for CH₄.The ratio of high-sensitivity channel and low-sensitivity channel are used to restrain errors caused by aerosol and surface albedo.In order to analyze the precision of the band ratio method,two retrieval methods are compared using surface albedo and aerosol error.One retrieval method is the optimal estimation method using whole band to retrieve CO₂ and CH₄.Another retrieval method is band-ratio method which is the improved optimal estimation method based on high-sensitivity channel and low-sensitivity channel.Based on the retrieval error analysis of aerosols of CO₂ and CH₄,band-ratio method is better than optimal estimation method.The band-ratio method can reduce the error by 2.5%and1.7%,respectively.The results show that the smaller boundary layer visibility can cause the greater influence on the retrieval error.The boundary layer visibility has most influence on retrieval error than the seasonal parameters and the boundary layer aerosol type.In the analysis of surface albedo error,the change of the surface albedo is5%,10%,20%and 30%.Results shows the band ratio method can restrain the retrieval error of CO₂ and CH₄ under 20%and satisfy the accuracy requirement of CO₂ and CH₄.Finally,we use GOSAT L3 data and three stations of TCCON retrieval date for verification.The time range is from April to September 2011.Results show Wollongong and Park Falls have better verification effects.The correlation coefficients for XCO₂ and XCH₄ are 0.83 and 0.75 at Park Falls between TCCON retrieval data and band ratio method.The correlation coefficients for XCO₂ and XCH₄are 0.9 and 0.68 at Park Falls between TCCON retrieval data and GOSAT L3 data.The correlation coefficients for XCO2 and XCH4 are 0.53 and 0.62 at Wollongong between TCCON retrieval data and band ratio method.The correlation coefficients for XCO₂ and XCH₄ are 0.45 and 0.86 at Wollongong between TCCON retrieval data and GOSAT L3 data.The correlation coefficients for XCO2 and XCH4 are 0.54 and 0.34at Bialystok between TCCON retrieval data and band ratio method.The correlation coefficients for XCO2 and XCH4 are 0.9 and 0.6 at Bialystok between TCCON retrieval data and GOSAT L3 data.The average standard deviation of the retrieval of XCO₂ of the daily data of TCCON and retrieval results from the ratio method used in this paper are between 0.38 and 0.75 ppm,indicating that the band ratio method is applicable and robust.The average standard deviation of the retrieval of XCO₂ of the daily data of TCCON and retrieval results from the ratio method used in this paper are between 0.38ppb and 0.75ppm,indicating that the ratio retrieval algorithm after band selection is well applicable and robust.The average standard deviation of the retrieval of XCH₄ of the daily data of TCCON and retrieval results from the ratio method used in this paper are between 2.3ppb and 5.4 ppb.The object of this study is to improve the retrieval accuracy of CO₂ and CH₄.This study has investigated the influencing factors that may have affected the retrieval of atmospheric CO₂ and CH₄ and established the criteria to channel selection using ratio retrieval method.This study provides some reference for the future study on quantitative measurement and analysis of greenhouse gases with higher accuracy.