Remote Sensing Estimation Of Dissolved Organic Carbon In Inland Lakes Based On The Fluorescence And Absorption Characteristics Of CDOM

Carbon is the foundation of all life on the earth.As the most important factor in the flow of energy and material in life activities,it plays a vital role in the global carbon cycle and the physical and chemical cycles of ecosystems.Although inland waters account for only about 1%of the global waters,the annual carbon buried quantity of inland lakes accounts for 25-58%of the ocean.The organic carbon pool of inland waters is mainly composed of dissolved organic carbon(DOC)and particulate organic carbon.The dissolved organic carbon is the largest organic carbon pool in inland lakes,and it is closely related to the photosynthesis and biological metabolic activities of phytoplankton in the water.Besides,DOC can effectively prevent ultraviolet from entering the underwater ecosystem and protect the underwater organisms due to the absorption of ultraviolet.However,there were still many scholars who believed that lakes only play a"pipeline"role in the global carbon cycle,and this phenomenon is the main reason for the current imbalance in the global carbon budget.Therefore,exploring the temporal and spatial evolution characteristics of dissolved organic carbon in the Yangtze River Delta(a typical region in China)using remote sensing technology is helping to provide scientific support for monitoring the pollution of inland lake organic,protecting drinking water sources,and providing data support for further improving the carbon cycle of inland waters.This paper analyzed the fluorescence fingerprint information of Colored dissolved organic matter(CDOM)of the lakes,interpreted the source and composition information of the dissolved organic carbon,explored the characteristics of fluorescence and optical of CDOM,and constructed optical indicators to distinguish the CDOM source.A DOC remote sensing estimation model based on different CDOM source was developed.And the lakes in the Yangtze River Delta were taken as the research object,the remote sensing estimation model of DOC is useful for real-time monitoring of DOC in inland lakes by exploring the temporal and spatial characteristics of DOC concentration in the Yangtze River Delta,revealing the driving mechanism of human activities and meteorological changes on the change of DOC concentration.It can provide a reference for the carbon cycle of regional lakes and even the global carbon cycle.The main research contents and conclusions of this thesis were as follows:(1)The fluorescence and absorption characteristics of CDOM in inland lakes were explored.CDOM absorption characteristics and fluorescence characteristics of inland lakes had obvious regional characteristics.The average value of CDOM specific absorption coefficient(a*350)in Lake Hongze and Lake Gaoyou was significantly higher than that of Lake Taihu and Lake Shijiu,indicating that the degree of CDOM humification in Lake Taihu was low,and the light absorption capacity of organic matter per unit concentration was low.Besides,fluorescence regional integration method(FRI)was used to quantitatively analyze CDOM source information and composition structure in different lakes.It was found that Humic acid-like components played a vital role in CDOM structure in the lakes,and it accounted for 27-53%of CDOM structure,with an average of 39%.The total contribution of endogenous components(sum of Tyrosine-like protein component,Tryptophan-like protein component,and soluble microbial by-product-like component)in Lake Taihu is 47%,which was much higher than that of other lakes,this phenomenon indicated algal blooms in eutrophic lakes degradation can change the information on the composition of organic matter in water.The biological source index(BIX)and the fluorescence index(FI)indicate that CDOM of inland lakes was characterized by the coexistence of terrestrial and endogenous sources.The principal component analysis(PCA)results showed that the humification parameter(HP)can effectively indicate the source information and composition structure changes of CDOM.(2)The optical response of CDOM fluorescence and absorption in inland lakes and the optical indicator to distinguish different CDOM sources were analyzed.There was a significant positive correlation between CDOM exogenous fluorescent component fulvic acid-like and CDOM absorption coefficient a_cdom254,indicating that CDOM absorption characteristics of lakes were mainly dominated by exogenous components,supplemented by endogenous components;HP had a significant positive correlation with the relative molecular mass parameter(M)and the specific absorption coefficient(a*350)(correlation coefficient(R)are 0.72 and 0.61,respectively),but the specific absorption coefficient a*350 is the ratio of the CDOM absorption coefficient to the DOC concentration,which is not suitable as an optical discriminant index.Therefore,the M value was used as an optical index to identify the source information of CDOM in water,and the inland lake was classified into two types:when the M value is greater than 8,it is a Type? water,and the CDOM is mainly input from endogenous source;when the M value is less than 8 It is a Type? water,and CDOM is mainly imported from exogenous source.(3)The changing pattern of CDOM absorption characteristics under different sources and the response of DOC were studied.The CDOM absorption spectrum of Type? water was significantly lower than that of Type?.The slope of CDOM spectrum S_275-295 of Type? water is higher than that of Type? water,indicating that Type? water was rich in macromolecular substances,with high aromatic concentration.The correlation between the CDOM absorption coefficient and DOC of two different water types gradually decreased with the increase of wavelength,and there was almost no correlation between DOC and CDOM absorption coefficient after 450nm.The correlation between the CDOM absorption coefficient of Type? water and DOC was significantly higher than that of Type? water,and there was no obvious correlation between the CDOM spectral slope S_275-295 and the relative molecular mass M value and DOC.(4)A DOC remote sensing estimation model based on fluorescence classification was developed.Based on the results of the classification of HP and M values,it was found that the Type? water caused by inputting of terrestrial sources had a high concentration of suspended particulate matter.And the remote sensing reflectance of Type? water and Type? water was significantly different in spectrum and magnitude,and a water classification strategy based on remote sensing reflectance to distinguish different CDOM sources had been established.An empirical model for DOC remote sensing estimation based on fluorescence classification and CDOM absorption coefficient was constructed,and it was evaluated through an independent measured data set.It had obtained better results than the random forest model and semi-analytical model.The validated UMRD(Unbiased median relative difference)was 17.35%,URMSPD(Unbiased root-mean-squared percentage difference)was 20.27%,and the Mean ratio was 1.02.Besides,the validation results of match-ups also showed that the DOC remote sensing estimation empirical model based on fluorescence classification has satisfactory accuracy,with an MRD(Median relative difference)of 19.92%,and an RMSPD(Root-mean-squared percentage difference)of 24.94%.The successful development of the DOC remote sensing estimation empirical model based on fluorescence classification provided a theoretical basis and technical support for the long-term series of remote sensing image-based DOC inversion and dynamic monitoring of lakes in the Yangtze River Delta in the future.(5)The spatial and temporal distribution pattern of DOC in the Yangtze River Delta and the driving factors of its change was revealed.Management Unit of the North Seas Mathematical Models(MUMM)atmospheric correction was carried out for the data set of 2016-2020 OLCI remote sensing image,and compared with the Satellite-situ Synchronization data,the accuracy result of MUMM atmospheric correction showed that the MRD is 29.81%and RMSPD is 32.79%,which provides data quality guarantee for estimating the spatial-temporal characteristics of DOC in the Yangtze River Delta.From remote sensing estimation results of DOC in lakes in the Yangtze River Delta,the spatial distribution of DOC concentration in lakes in the Yangtze River Delta had obvious regional characteristics.The average DOC concentration of 21 lakes in the Yangtze River Delta was 3.43 mg/L,and the DOC concentration of lakes in the Huai River Basin was higher than that of the lakes in the middle and lower reaches of the Yangtze River.The high concentration of DOC in most lakes was distributed in the rivers entering the lake,indicating that external source input is the main reason for the increase of DOC in lakes.The ensemble empirical mode decomposition method(EEMD)was selected to explore the monthly variation of DOC in the Yangtze River Delta.It was found that the DOC in the Yangtze River Delta shows a trend of first decreasing and then increasing.The concentration of DOC in summer was lower,and the concentration of DOC was higher in autumn and winter.Meanwhile,the DOC value in the Yangtze River Delta mainly had a year cyclical change,and the DOC in the Qiantangjiang Basin was mainly cyclical and half-year.Analyzing the influence of natural factors,namely meteorological factors,on lake DOC in the Yangtze River Delta,the results showed that the DOC concentration that varies annually was mainly affected by high temperature and rain(temperature and rainfall),followed by wind speed.