| Carbon dioxide(CO2)is the most important greenhouse gas in the atmosphere and it has an important impact on global climate change.As a resident gas,the concentration and concentration gradient of CO2 in the atmosphere are in very low level.Therefore,high-precision detection of CO2 concentration in the atmosphere is vital for studying the change law of CO2 source and sink and predicting global climate change.Although the CO2 ground observation accuracy is high,sparse distribution of observation sites makes it difficult to meet the application needs.Satellite remote sensing,characterized by global coverage and high frequency sampling,makes global observation and quick data update possible,so it can better meet the application needs in related fields.Therefore,in recent years,the CO2 satellite remote sensing detection technology has witnessed rapid development.At present,many satellites are already in operation and many others are to be lauched in the world for the purpose of detecting CO2 and other greenhouse gases.In order to obtain greenhouse observation data that can be applied in practice,high precision atmospheric CO2 satellite remote sensing is vital.How to achieve high-precision atmospheric CO2 inversion has become a key issue in this remote sensing technology.Cloud pollution problem in remote sensing data,surface and atmospheric parameters are difficult to obtain accurately,the presence and changes of chlorophyll fluorescence in vegetation,etc.,all interfere with the inversion of CO2,affecting the accuracy of CO2 inversion.Therefore,how to overcome the influence of interference factors has become a key issue for atmospheric CO2 inversion.Statistically speaking,more than 70%of the world’s regions are covered by clouds.Therefore,the problem of remote sensing data cloud pollution has become the first problem to be solved in CO2 inversion.This paper firstly studies the cloud detection method,through the inversion of CO2 synchronous observation of O2 A-band data,combined with statistical methods,cloud detection and cloud pixel filtering are implemented in a threshold manner.The cloud detection algorithm is verified by the Greenhouse Gases Monitoring Instrument(GMI)data.Compared with the MODIS(Moderate Resolution Imaging Spectradiometer)cloud judgment product,the results show that the cloud screening algorithm can screen out 90%of the clear sky pixels.Atmospheric scattering changes the distribution of light radiation and the photon transmission path,affecting the quantitative characterization of atmospheric CO2 radiation transmission.This has become an important source of CO2 inversion error,which seriously affects the inversion accuracy of atmospheric CO2 satellite remote sensing.In the course of this study,it was found that the content of oxygen in the atmosphere is highly stable.In the study of atmospheric CO2 inversion methods,this characteristic of oxygen can be used for atmospheric scattering correction.However,there is vegetation chlorophyll fluorescence in the O2 A absorption zone,which is a kind of light radiation emitted by the vegetation itself.It is not the direct solar spectrum radiation based on atmospheric CO2 remote sensing.Therefore,the presence of vegetation chlorophyll fluorescence affects the reaction of oxygen to CO2.In the remote sensing of greenhouse gas satellites,the chlorophyll fluorescence satellite observation has a single information source and it is difficult to stably unwind the influence of atmospheric scattering in the O2 A band.The inversion method has different effects on CO2 inversion.Based on the investigation of global vegetation fluorescence distribution,the effects of fluorescence on atmospheric CO2 column content(XCO2)were simulated by PPDF method and full physical inversion method.The results show that for the PPDF method,if the influence of aerosol is not considered,the fluorescence intensity increases from 0.1 to 1.8(mW·m-2·sr-1·nm-1),which will result in 0.1 to 2 ppm of CO2 inversion results.Deviation;if considering the influence of aerosol and surface reflectivity,it will cause a deviation of 0.1-3 for the inversion result of CO2.For the full physical inversion method,when the influence of fluorescence is neglected,in extreme cases,an overestimation of 15 ppm of CO2 can be caused.In view of the spectral distribution characteristics of remote sensing data,in the remote sensing data inversion experiment,the PPDF method is used to correct the influence of fluorescence on the GMI spectrum of the GF-5 satellite.Based on the research of key problems such as cloud detection and oxygen correction,this paper studies the integrity of atmospheric CO2 inversion method based on the characteristics of high-resolution satellite remote sensing data,and carries out preliminary experimental verification with its remote sensing data.The inversion results are compared with the TCCON,GOSAT and OCO-2 CO2 inversion results.The results show that the average accuracy of CO2 inversion is better than 1%.Among them,the correlation between the inversion results and the results of the TCCON site is compared,showing a correlation coefficient of 0.85,Due to the influence of vegetation chlorophyll fluorescence on the Park Falls and Sodankyla sites,the GMI inversion results near the site is higher than those of TCCON.After correcting the effect of chlorophyll fluorescence on vegetation,the correlation coefficient between GMI and CO2 inversion results at TCCON site increases to 0.913. |
PDF Full Text Request