Retrieval Of Vegetation Chlorophyll Fluorescence Based On Greenhouse Gas Satellite Remote Sensing Data

Chlorophyll fluorescence is produced during vegetation photosynthesis,which is closely related to photochemical action,and is an important index to reflect the biochemical process of vegetation.The difficult task of the remote sensing retrieval of chlorophyll fluorescence is caused by its weak signal,in which only about 2%of the absorbed light is converted into fluorescence.Greenhouse gas is an important factor leading to global climate change,which has a important impact on the production and life of human society.Many remote sensing satellites dedicated to greenhouse gases monitoring have been developed in recent years,for instance,GOSAT of Japan,OCO2 of Unite States of America,GF-5 of China,etc.whose basic features include oxygen channels and extremely high spectral resolution.The chlorophyll fluorescence band,which partially overlaps with the oxygen absorption band of carbon monitoring satellites is located between 650 nm and 800nm,thus capable of retrieving chlorophyll fluorescence for carbon monitoring satellites.It is of great prospects to the combination of vegetation chlorophyll fluorescence and greenhouse gas monitoring,since vegetation biochemical processes are important sinks of greenhouse gases.With the bases of using OCO-2 satellite data to retrieve chlorophyll fluorescence,this thesis explores the possibility of retrieving chlorophyll fluorescence from GMI data on GF-5 satellite,in order to expand the application in vegetation chlorophyll fluorescence monitoring based on GMI greenhouse gas monitoring ability.Chlorophyll fluorescence radiation signal is weak.How to peel off the influence of surface and atmosphere in the process of remote sensing is the key to obtaining good inversion results.Based on studying surface reflection,atmospheric radiation and absorption,and solar spectral lines,two technical routes are adopted in the retrieving methods:the Fraunhofer line methods(FLD)with very narrow linewidth and oxygen absorption line with plenty of spectral lines in this thesis.Based on the extremely narrow Fraunhofer line method,the principle is that the characteristics of surface reflection and atmospheric radiation on both sides of its spectral line are relatively consistent,which is called FLD method.Further consideration in detail of the possible changes in linear and nonlinear on both sides of Fraunhofer line,which expand the methods based of Fraunhofer line,namely the modified FLD(3FLD)and the improved FLD(iFLD).The principle of plentiful oxygen absorption lines is to use more spectral lines to overcome the disadvantage of noise in remote sensing data,which is the SFM method.To study the ability of retrieving chlorophyll fluorescence under different vegetation cover situations,three areas,namely the Amazon forest,African grassland,were selected in the retrieval experiment on the principle of high,medium,and low fluorescence intensity.We use FLD,3FLD,and iFLD methods,respectively for the retrieval,which is based on the KI Fraunhofer line near 770.1 nm.It is shown in the retrieval results that the several methods are highly consistent with the fluorescence products provided by OCO-2,with the degree of consistency from high to low is iFLD,3FLD,and FLD.The SFM method using oxygen absorption line for retrieval has the same high consistency with OCO-2 fluorescence products,and it is higher than the FLD series methods.Specifically,the retrieval results show that the stronger of fluorescence,the higher the retrieval accuracy,whereas the weaker the fluorescence,the lower the retrieval accuracy.In January 2019,the chlorophyll fluorescence spectrum in the Amazon forest area during the rainy season was strong,and the SFM retrieval method had the highest accuracy,with the correlation coefficient with OCO-2 products reaching 0.934,and the correlation coefficient of iFLD retrieval results in the same period was 0.882.In July 2019 when the chlorophyll fluorescence was weak in the desert area,the retrieval accuracy of SFM method is low,with the correlation coefficient also reaches 0.887,and the correlation coefficient of iFLD inversion results in the same period was 0.83 in the same period.The GMI sensor onboard the GF-5 satellite is developed by the author’s department.It is an important goal of this thesis to carry out the retrieval of vegetation chlorophyll fluorescence and expand the application capability of GMI data based on the GMI data of GF-5.Since the band range of GMI is slightly narrower than that of OCO-2,there is no suitable Fraunhofer line required for the FLD method.Therefore,the SFM retrieval method is used in GMI data.Amazon forest,African grassland,and Sahara desert are still selected as retrieval experimental areas.There are some differences between the results based on GMI remote sensing data and the results based on OCO-2 remote sensing data,taken the Amazon forest with strong fluorescence as an example,in January 2019,the retrieval results of GMI and OCO are 0.8103 w·m-2·sr1·nm-1 and 0.8029 w·m-2·sr-1·nm-1,while the value of OCO-2 fluorescence product is 0.7632 w·m-2·sr-1·nm-1,which shows that the GMI retrieval result deviates more from the OCO-2 fluorescence product than the OCO-2 retrieval result.This phenomenon also exists in other experimental areas and the data retrieval results in July 2019.Generally,the use of GMI remote sensing data receives good results in the retrieval of vegetation fluorescence intensity,however,the deviation of GMI data retrieval results is slightly larger than that of OCO-2,which is probably caused by the inconstancy in the location,environment,and observation conditions of the measuring points of GMI and OCO-2,or both,which there will be more demonstration and research in further study,to realize the retrieval and application of vegetation fluorescence of GMI data with high quality.